Rohit Verma

PUBLICATIONS

PUBLICATIONS LIST

Selective Graph Convolutional Network for Efficient Routing

Fourth International Conference on AI-MLSystems

Conferences K Nikita, Hayagreev J, Rohit Verma, Rajeev Shorey

Selective Graph Convolutional Network for Efficient Routing

K Nikita, Hayagreev J, Rohit Verma, Rajeev Shorey Conferences

Modern-day networks require highly intelligent and dynamic routing policies that are capable of routing data efficiently over a wide range of changing topology and load conditions. Machine learning frameworks like graph neural networks and multi-agent reinforcement learning have been applied to routing scenarios that deal with heterogeneous graphs. However, these solutions cannot sustain optimality under higher packet loads because deep Q networks and graph neural networks for a single agent indiscriminately factor in all neighbouring agents during the training phase. This also leads to higher computational costs during training and implementation.
In this paper, we propose a Selective Graph Convolutional Network (Sel_GCN ) model using Multi-Agent Deep Q Networks and Attention to maximize throughput and minimize latency over various packet load conditions. We evaluate our technique on a simulator and compare it to previous implementations of reinforcement learning and graph neural networks. Our results show that Sel_GCN achieves 100% packet delivery while reducing the CPU and memory usage by about 50% and 45% respectively.

On-the-Go Automated Break Recommendation for Stress Avoidance
during Highway Driving

ACM Journal on Autonomous Transportation Systems

Journal Paper Rohit Verma, Bivas Mitra, Sandip Chakraborty

On-the-Go Automated Break Recommendation for Stress Avoidance during Highway Driving

Rohit Verma, Bivas Mitra, Sandip Chakraborty Journal Paper

Continuous cab driving is considered one of the highly stressful jobs, although the drivers ignore that many a time. Taking a break from manual driving or transferring the control to another driver to release the stress would be an easy, intuitive solution, although the challenge is to detect the driving stress while the trip is going on. As driving stress depends on multiple diverse environmental and affective features, we, in this paper, develop a novel assistive system, SmartHalt, which continuously senses the driving environment (like road type, congestion, driver's driving pattern, etc.) and then utilizes a spatial time series model of the driving environment with a deep learning framework to predict whether the driver will be stressed while on the trip. The model also considers the personality traits of the drivers along with the spatio-temporal features to differentiate the impact of stress on the driving behaviour for different drivers and recommends taking a break soon before the driving behaviour drops below a critical level. A thorough analysis of the model over 7 different drivers for a 10 month-long experiment over 204871 km of driving data reveals that the proposed approach can significantly improve driving behaviour by recommending a driving break at proper times. Following the recommendation by SmartHalt improves the driving score by ≈ 50% and reduces the number of driving offences by ≈ 50%. SmartHalt can help develop advanced driving assisting system (ADAS) platforms that understand the affective states of the driver and thus can be helpful for semi-autonomous driving environments for effective driver-vehicle interactions.

Towards a Flexible Scale-out Framework for Efficient Visual Data Query Processing

arXiv

Preprint PDF Rohit Verma, Arun Raghunath

Towards a Flexible Scale-out Framework for Efficient Visual Data Query Processing

Rohit Verma, Arun Raghunath Preprint Paper

There is growing interest in visual data management systems that support queries with specialized operations ranging from resizing an image to running complex machine learning models. With a plethora of such operations, the basic need to receive query responses in minimal time takes a hit, especially when the client desires to run multiple such operations in a single query. Existing systems provide an ad-hoc approach where different solutions are clubbed together to provide an end-to-end visual data management system. Unlike such solutions, the Visual Data Management System (VDMS) natively executes queries with multiple operations, thus providing an end-to-end solution. However, a fixed subset of native operations and a synchronous threading architecture limit its generality and scalability.
In this paper, we develop VDMS-Async that adds the capability to run user-defined operations with VDMS and execute operations within a query on a remote server. VDMS-Async utilizes an event-driven architecture to create an efficient pipeline for executing operations within a query. Our experiments have shown that VDMS-Async reduces the query execution time by 2-3X compared to existing state-of-the-art systems. Further, remote operations coupled with an event-driven architecture enables VDMS-Async to scale query execution time linearly with the addition of every new remote server. We demonstrate a 64X reduction in query execution time when adding 64 remote servers.

Towards an Efficient Federated Learning Framework with Selective Aggregation

Bangalore, India

Proceedings of the 16th International Conference on COMmunication Systems and NETworkS (COMSNETS 2024)

Conferences Anirudha Kulkarni, Abhinav Kumar, Rajeev Shorey, Rohit Verma

Towards an Efficient Federated Learning Framework with Selective Aggregation

Anirudha Kulkarni, Abhinav Kumar, Rajeev Shorey, Rohit Verma Conferences

Federated Learning shows promise for collaborative, decentralized machine learning but faces efficiency challenges; primarily network straggler-induced latency bottlenecks and the need for complex aggregation techniques. To address these issues, ongoing research explores asynchronous FL, i.e., federated learning models, including an Asynchronous Parallel Federated Learning framework. This study investigates the impact of varying worker node numbers on key metrics. More nodes offer faster convergence but may increase communication overhead and straggler vulnerability. We aim to quantify how the number of worker node variations for one global aggregation can affect convergence speed, communication efficiency, model accuracy, and system robustness, optimizing asynchronous FL system configurations. This work is crucial for practical and scalable FL applications, mitigating network stragglers, data distribution, and security challenges. This work introduces Asynchronous Parallel Federated Learning, showcasing a paradigm shift by selectively aggregating early arriving worker node updates with a novel parameter `x', improving efficiency and reshaping FL applications.

FedNSE: Optimal Node Selection for Federated Learning with
Non-IID Data

Bangalore, India

Proceedings of the 15th International Conference on COMmunication Systems and NETworkS (COMSNETS 2023)

Conferences Sourav Bansal, Manav Bansal, Rohit Verma, Rajeev Shorey and Huzur Saran

FedNSE: Optimal Node Selection for Federated Learning with Non-IID Data

Sourav Bansal, Manav Bansal, Rohit Verma, Rajeev Shorey and Huzur Saran Conferences

Federated Learning relies heavily on the data available at the worker nodes. In a majority of practical use-cases, the data at the worker nodes are Non-IID (non-independent-and-identically-distributed) in nature. The central server often only selects a subset of the worker nodes to compute the global model. If the subset has more nodes with heterogeneous data, the overall performance would be compromised. Selecting an optimal subset of worker nodes from the available worker nodes is crucial for improving the accuracy and convergence time of Federated Learning models. In this paper, we analyse the impact of Non-IID data, measured using quantity skewness, on the performance of the federated learning model and design approaches to detect the level of skewness of the data distribution at the worker nodes. Following this, we design an approach Federated Node Selection with Entropy (FedNSE) to select the optimal set of worker nodes to improve the accuracy and convergence time of a federated learning model. Our experiments performed in a simulated Federated Learning environment show that using FedNSE that selects the optimal subset of worker nodes improves the convergence time and accuracy compared to the existing approaches.

Multi-Agent Packet Routing (MAPR): Co-Operative Packet Routing Algorithm with Multi-Agent Reinforcement Learning

Bangalore, India

Proceedings of the 15th International Conference on COMmunication Systems and NETworkS (COMSNETS 2023)

Conferences Aniket Modi, Rishi Shah, Krishnanshu Jain, Rohit Verma, Rajeev Shorey and Huzur Saran

Multi-Agent Packet Routing (MAPR): Co-Operative Packet Routing Algorithm with Multi-Agent Reinforcement Learning

Aniket Modi, Rishi Shah, Krishnanshu Jain, Rohit Verma, Rajeev Shorey and Huzur Saran Conferences

Packet Routing has been investigated intensively in computer networking with numerous existing solutions. However, with the increasing usage of distributed networks that have frequent changes in the network architecture and the environment, there is a need of solutions that learn changes in the network and modify the routing policy in real time. Multi-Agent Reinforcement Learning (MARL) is emerging as a promising approach for such distributed networks. But, existing MARL based approaches are limited in the level of co-operation amongst agents and use Deep Q-Networks (DQNs) that could easily lead to overestimation bias. Moreover, for a co-operative learning scenario such as MARL, it is crucial for the reward to be generated in a way that agents that could hamper the network performance are not created.
In this paper, we design a packet routing algorithm Multi-Agent Packet Routing (MAPR) that ensures that the neighbouring agent information is captured temporally throughout the training phase. MAPR uses a Double Deep Q Network (DDQN) to ensure that the overestimation bias is avoided in the learning phase. Furthermore, we design a delay-tolerant global-cum-local reward approach to ensure that performance hampering agents are not created. We evaluate the system using a Python-based simulator that has been designed over multiple network topologies to show the effectiveness of the algorithm.

DeepDish on a diet: Low-Latency, Energy-Eﬀicient Object-Detection and Tracking at the Edge

Proceedings of the 5th International Workshop on Edge Systems, Analytics and Networking

Conferences Matthew Danish, Rohit Verma, Justas Brazauskas, Ian Lewis and Richard Mortier

DeepDish on a diet: Low-Latency, Energy-Eﬀicient Object-Detection and Tracking at the Edge

Matthew Danish, Rohit Verma, Justas Brazauskas, Ian Lewis and Richard Mortier Conferences

Intelligent sensors using deep learning to comprehend video streams have become commonly used to track and analyse the movement of people and vehicles in public spaces. The models and hardware become more powerful at regular and frequent intervals. However, this computational marvel has come at the expense of heavy energy usage. If intelligent sensors are to become ubiquitous, such as being installed at every junction and frequently along every street in a city, then their power draw will become non-trivial, posing a severe downside to their usage. We explore Multi-Object Tracking (MOT) solutions based on our custom system that use less power while still maintaining reasonable accuracy.

Impact of Driving Behavior on Commuter’s Comfort during Cab Rides: Towards a New Perspective of Driver Rating

ACM Transactions on Intelligent Systems and Technology

Journal Paper Rohit Verma, Sugandh Pargal, Debasree Das, Tanushree Parbat, S. Sai Kambalapalli, Bivas Mitra, Sandip Chakraborty

Impact of Driving Behavior on Commuter’s Comfort during Cab Rides: Towards a New Perspective of Driver Rating

Rohit Verma, Sugandh Pargal, Debasree Das, Tanushree Parbat, S. Sai Kambalapalli, Bivas Mitra, Sandip Chakraborty Journal Paper

Commuter comfort in cab rides affects driver rating as well as the reputation of ride-hailing firms like Uber/Lyft. Existing research has revealed that commuter comfort not only varies at a personalized level but also is perceived differently on different trips for the same commuter. Furthermore, there are several factors, including driving behavior and driving environment, affecting the perception of comfort. Automatically extracting the perceived comfort level of a commuter due to the impact of the driving behavior is crucial for a timely feedback to the drivers, which can help them to meet the commuter's satisfaction. In light of this, we surveyed around 200 commuters who usually take such cab rides and obtained a set of features that impact comfort during cab rides. Following this, we develop a system Ridergo which collects smartphone sensor data from a commuter, extracts the spatial time series feature from the data, and then computes the level of commuter comfort on a five-point scale with respect to the driving. Ridergo uses a Hierarchical Temporal Memory model-based approach to observe anomalies in the feature distribution and then trains a Multi-task learning-based neural network model to obtain the comfort level of the commuter at a personalized level. The model also intelligently queries the commuter to add new data points to the available dataset and, in turn, improve itself over periodic training. Evaluation of Ridergo on 30 participants shows that the system could provide efficient comfort score with high accuracy when the driving impacts the perceived comfort.

FedFM: Towards a Robust Federated Learning Approach For
Fault Mitigation at the Edge Nodes

Bangalore, India

Proceedings of the 14th International Conference on COMmunication Systems and NETworkS (COMSNETS 2022)

Conferences PDF Manupriya Gupta, Pawas Goyal, Rohit Verma, Rajeev Shorey, and Huzur Saran

FedFM: Towards a Robust Federated Learning Approach For Fault Mitigation at the Edge Nodes

Manupriya Gupta, Pawas Goyal, Rohit Verma, Rajeev Shorey, and Huzur Saran Conferences

Federated Learning deviates from the norm of "send data to model" to "send model to data". When used in an edge ecosystem, numerous heterogeneous edge devices collecting data through different means and connected through different network channels get involved in the training process. Failure of edge devices in such an ecosystem due to device fault or network issues is highly likely. In this paper, we first analyse the impact of the number of edge devices on an FL model and provide a strategy to select an optimal number of devices that would contribute to the model. We observe how the edge ecosystem behaves when the selected devices fail and provide a mitigation strategy to ensure a robust Federated Learning technique.

SmartSplit: Latency-Energy-Memory Optimisation for CNN Splitting on Smartphone Environment

Bangalore, India

Proceedings of the 14th International Conference on COMmunication Systems and NETworkS (COMSNETS 2022)

Conferences PDF Ishaan Prakash, Aniruddh Bansal, Rohit Verma, and Rajeev Shorey

SmartSplit: Latency-Energy-Memory Optimisation for CNN Splitting on Smartphone Environment

Ishaan Prakash, Aniruddh Bansal, Rohit Verma, and Rajeev Shorey Conferences

Artificial Intelligence has now taken centre stage in the smartphone industry owing to the need of bringing all processing close to the user and addressing privacy concerns. Convolution Neural Networks (CNNs), which are used by several AI applications, are highly resource and computation intensive. Although new generation smartphones come with AI-enabled chips, minimal memory and energy utilisation is essential as many applications are run concurrently on a smartphone. In light of this, optimising the workload on the smartphone by offloading a part of the processing to a cloud server is an important direction of research. In this paper, we analyse the feasibility of splitting CNNs between smartphones and cloud server by formulating a multi-objective optimisation problem that optimises the end-to-end latency, memory utilisation, and energy consumption. We design SmartSplit, a Genetic Algorithm with decision analysis based approach to solve the optimisation problem. Our experiments run with multiple CNN models show that splitting a CNN between a smartphone and a cloud server is feasible. The proposed approach, SmartSplit fares better when compared to other state-of-the-art approaches.

Latency-Memory Optimized Splitting of Convolution Neural Networks for Resource Constrained Edge Devices

Bangalore, India

Proceedings of the 14th International Conference on COMmunication Systems and NETworkS (COMSNETS 2022)

Conferences PDF Tanmay Jain, Avaneesh, Rohit Verma, and Rajeev Shorey

Latency-Memory Optimized Splitting of Convolution Neural Networks for Resource Constrained Edge Devices

Tanmay Jain, Avaneesh, Rohit Verma, and Rajeev Shorey Conferences

With the increasing reliance of users on smart devices, bringing essential computation at the edge has become a crucial requirement for any type of business. Many such computations utilize Convolution Neural Networks (CNNs) to perform AI tasks, having high resource and computation requirements, that are infeasible for edge devices. Splitting the CNN architecture to perform part of the computation on edge and remaining on the cloud is an area of research that has seen increasing interest in the field. In this paper, we assert that running CNNs between an edge device and the cloud is synonymous with solving a resource-constrained optimization problem that minimizes the latency and maximizes resource utilization at the edge. We formulate a multi-objective optimization problem and propose the LMOS algorithm to achieve a Pareto efficient solution. Experiments done on real-world edge devices show that LMOS ensures feasible execution of different CNN models at the edge and also improves upon existing state-of-the-art approaches.

Real-Time Data Visualisation on the Adaptive City Platform

Cambridge, UK

Proceedings of the 8th ACM International Conference on Systems for Energy-Efficient Built Environments (BuildSys 2021)

Conferences PDF Justas Brazauskas, Rohit Verma, Vadim Safronov, Matthew Danish, Ian Lewis, Richard Mortier

Real-Time Data Visualisation on the Adaptive City Platform

Justas Brazauskas, Rohit Verma, Vadim Safronov, Matthew Danish, Ian Lewis, Richard Mortier Conferences

In smart buildings research, the integration of Building Information Models (BIM), Building Management Systems (BMS), and Internet of Things (IoT) is of paramount importance. However, such integration often overlooks real-time building data visualisation. In this demo, we examine challenges related to spatiotemporal data representation and novel visualisation methods in smart environments. Following this, we present the front-end design of our Adaptive City Platform (ACP), a system for collecting, processing and visualising building information and sensor data in real-time.

Do we want the New Old Internet? Towards Seamless and Protocol-Independent IoT Application Interoperability

Cambridge, UK

Proceedings of the 20th ACM Workshop on Hot Topics in Networks (HotNets 2021)

Conferences PDF Vadim Safronov, Justas Brazauskas, Matthew Danish, Rohit Verma, Ian Lewis, Richard Mortier

Do we want the New Old Internet? Towards Seamless and Protocol-Independent IoT Application Interoperability

Vadim Safronov, Justas Brazauskas, Matthew Danish, Rohit Verma, Ian Lewis, Richard Mortier Conferences

IoT is developing rapidly with frequently appearing new wireless standards and applications. However, besides a large number of IoT benefits, its further development is now being slowed down due to the repetition of old Internet development flaws while dealing with IoT heterogeneity. The current misleading trend aims to solve all IoT interoperation problems by inserting IP Addresses into those wireless protocols where the IP stack clearly slows down application performance and drains the battery, e.g. LPWANs such as LoRaWAN and SigFox. This paper tackles IoT heterogeneity from a different perspective: it is the application interoperation which matters the most. The protocols beneath the application layer shall work for smooth upper-layer service provisioning where the IP shall be just one of the many underlying integration options instead of being the essential one. Inspired by previous proposals for a more flexible internetworking architecture, this paper applies those theoretical concepts in practice by proposing a protocol-independent distributed interoperation model for smooth service provisioning over heterogeneous IoT wireless contexts. The arguments pro the IP-agnostic IoT application interoperation are supported by the model's prototype which showed 1.6-2 times faster MQTT application operation over LoRa and WiFi compared to the legacy IP-based MQTT provisioning over that protocols.

RACER: Real-Time Automated Complex Event Recognition in Smart Environments

Beijing, China

Proceedings of the 29th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (SIGSPATIAL 2021)

Conferences PDF Rohit Verma, Justas Brazauskas, Vadim Safronov, Matthew Danish, Ian Lewis, Richard Mortier

RACER: Real-Time Automated Complex Event Recognition in Smart Environments

Rohit Verma, Justas Brazauskas, Vadim Safronov, Matthew Danish, Ian Lewis, Richard Mortier Conferences

As smart environments become laden with more and more sensors, there has been a need to develop systems that could derive useful information from these sensors and make the smart environments smarter. Complex Event Processing (CEP) has emerged as a popular strategy to identify crucial events from sensor data. However, the existing CEP strategies overlook the relationship with other sensors in the spatial vicinity and understate the temporal variation of sensor data. In this paper, we develop RACER, which is an end-to-end complex event processing system that takes into consideration both the spatial location of the sensor in observation and the varying impact of temporal changes in the sensor data. Experiments performed for a duration of five months over both collected and live streaming data shows that RACER fares well compared to the other state-of-the-art approaches.

Data Management for Building Information Modelling in a Real-Time Adaptive City Platform

arXiv

Preprint PDF Justas Brazauskas, Rohit Verma, Vadim Safronov, Matthew Danish, Jorge Merino, Xiang Xie, Ian Lewis, Richard Mortier

Data Management for Building Information Modelling in a Real-Time Adaptive City Platform

Justas Brazauskas, Rohit Verma, Vadim Safronov, Matthew Danish, Jorge Merino, Xiang Xie, Ian Lewis, Richard Mortier Preprint Paper

Legacy Building Information Modelling (BIM) systems are not designed to process the high-volume, high-velocity data emitted by in-building Internet-of-Things (IoT) sensors. Historical lack of consideration for the real-time nature of such data means that outputs from such BIM systems typically lack the timeliness necessary for enacting decisions as a result of patterns emerging in the sensor data. Similarly, as sensors are increasingly deployed in buildings, antiquated Building Management Systems (BMSs) struggle to maintain functionality as interoperability challenges increase. In combination these motivate the need to fill an important gap in smart buildings research, to enable faster adoption of these technologies, by combining BIM, BMS and sensor data. This paper describes the data architecture of the Adaptive City Platform, designed to address these combined requirements by enabling integrated BIM and real-time sensor data analysis across both time and space.

SenseRT: A Streaming Architecture for Smart Building Sensors

arXiv

Preprint PDF Rohit Verma, Justas Brazauskas, Vadim Safronov, Matthew Danish, Jorge Merino, Xiang Xie, Ian Lewis, Richard Mortier

SenseRT: A Streaming Architecture for Smart Building Sensors

Rohit Verma, Justas Brazauskas, Vadim Safronov, Matthew Danish, Jorge Merino, Xiang Xie, Ian Lewis, Richard Mortier Preprint

Building Management Systems (BMSs) have evolved in recent years, in ways that require changes to existing network architectures that follow the store-then-analyse approach. The primary cause is the increasing deployment of a diverse range of cost-effective sensors and actuators in smart buildings that generate real-time streaming data. Any in-building system with a large number of sensors needs a framework for real-time data collection and concurrent stream processing from sensors connected using a range of networks.
We present SenseRT, a system for managing and analysing in-building real-time streams of sensor data. SenseRT collects streams of real-time data from sensors connected using a range of network protocols. It supports concurrent modules simultaneously performing stream processing over real-time data, asynchronously and non-blocking, with results made available with minimal latency. We describe a prototype implementation deployed in two University department buildings, demonstrating its effectiveness.

Smartphones for Public Transport Planning and Recommendation in Developing Countries - A Review

WIREs Data Mining and Knowledge Discovery

Journal Paper PDF Rohit Verma, Sandip Chakraborty

Smartphones for Public Transport Planning and Recommendation in Developing Countries - A Review

Rohit Verma, Sandip Chakraborty Journal Paper

In this era of connected systems that have penetrated everywhere, transport units have become a significant source of data, collected from commuters, vehicles, drivers, or any section being touched by the transport system. This data, which has both spatial as well as temporal aspects, is utilized for a plethora of services like travel assistant systems, multi-modal transport solutions, real-time travel information, smart parking, autonomous vehicles, to name a few. With the current buzz of sustainable transport, the use of public transport systems have been popularized owing to the economic and environmental savings. In this review paper, we provide a highlight of works which have tried to utilize techniques to improve multiple sections of the public transport system, primarily focusing on developing economies, thus improving the overall commute experience at various countries.

A Smartphone-based Passenger Assistant for Public Bus Commute in Developing Countries

IEEE Transactions on Computational Social Systems (2020)

Journal Paper PDF Rohit Verma, Aviral Shrivastava, Kingshuk De, Bivas Mitra, Sujoy Saha, Niloy Ganguly, Subrata Nandi and Sandip Chakraborty

A Smartphone-based Passenger Assistant for Public Bus Commute in Developing Countries

Rohit Verma, Aviral Shrivastava, Kingshuk De, Bivas Mitra, Sujoy Saha, Niloy Ganguly, Subrata Nandi and Sandip Chakraborty Journal Paper

Although public transports vehicles like buses have always been a cheap means of commuting in the cities of many developing countries, it is always considered as a secondary mode of transport owing to poor infrastructure, chaotic and reckless driving habits and absence of any proper information system in buses. Based on rigorous experiments carried over a period of two years and multiple surveys, we have tried to learn the problems faced by the bus commuters. As a solution, in this work, we develop a novel energy efficient system which would help commuters navigate through their journey safely. Along with making them aware of any upcoming points of concerns (PoC) like sudden bumps, sharp turns, bad roads etc, we also inform commuters about the expected time of arrival at the destination. The system makes use of several landmarks like speed breakers, turns and bus stops on a trail stored in a specialized data structure, the probabilistic timed automata. We conducted extensive experiments using 25 volunteers over 50 trails. The system showed an average localization error of only 50m and mean ETA error of 2.5 minutes and a fairly high alert prediction accuracy, while consuming significantly less energy when compared to GPS.

Avoiding Stress Driving: Online Trip Recommendation from Driving Behavior Prediction

KYOTO - JAPAN

Proceedings of the 2019 IEEE International Conference on Pervasive Computing and Communications (PerCom 2019)

Conferences PDF PPT Rohit Verma, Bivas Mitra and Sandip Chakraborty

Avoiding Stress Driving: Online Trip Recommendation from Driving Behavior Prediction

Rohit Verma, Bivas Mitra and Sandip Chakraborty Conferences

The growth in the market for cab companies like Uber has opened the door to high-income options for drivers. However, in order to boost their income, drivers many a time resort to accepting trips which increases their stress resulting in poor driving quality and accidents in serious cases. Every driver handles stress differently and the trip recommendation thus needs to be on a personalized level. In this paper, we explore historical trip data to compute the driving stress and its impact on various driving behavioral features, captured through vehicle-mounted GPS and inertial sensors. We utilize a Multi-task Learning based Neural Network model to learn both the common features and the personalized features from the driving data to predict the stress level of a driver. We further establish a causal relationship between the stress level of a driver and his driving behavior. Finally, we develop a trip recommendation system for cab drivers to avoid stress driving. The models have been tested over both a publicly available dataset with 6 drivers for 500 minutes of driving data and an in-house collected dataset from 8 drivers over 1700 trips for 5 months. We observe that the proposed model gives an average prediction accuracy of 94% with low false-positive rates. We also observed that the driving behavior is improved when a driver takes a recommended trip.

Mining Spatio-temporal Data for Computing Driver Stress and
Observing Its Effects on Driving Behavior

SEATTLE - USA

Proceedings of the 26th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (SIGSPATIAL 2018)

Conferences PDF Rohit Verma, Gyanesh Prajjwal, Bivas Mitra and Sandip Chakraborty

Mining Spatio-temporal Data for Computing Driver Stress and Observing Its Effects on Driving Behavior

Rohit Verma, Gyanesh Prajjwal, Bivas Mitra and Sandip Chakraborty Conferences

With the increase in road fatalities due to various factors like aggressive driving and road rage, quantifying and monitoring the stress level of a driver is an important task for the preparation of driving rosters for the cab companies. Stress monitoring using physiological sensors is a costly and obstructive task, while stress factors impact dierently for dierent individuals based on their personality traits. In this paper, we develop a learning-based model to predict the stress level of a driver and its eect on his driving behavior, solely based on spatio-temporal driving data collected through GPS and inertial sensors. We further establish a correlation between the stress level of a driver and his driving behavior; thus, we develop a complete system to infer stress proling and its impact on driving behavior based on spatio-temporal driving data. e model has been tested over a publicly available dataset with 6 drivers for 500 minutes of driving data. We observe that the proposed model gives an average prediction accuracy of 79% with low false-positive rates.

ComfRide:A Smartphone based System for Comfortable Public Transport Recommendation

VANCOUVER - CANADA

Proceedings of the 12th ACM Conference on Recommender Systems (RecSys 2018)
Project page Link

Conferences PDF PPT Rohit Verma, Surjya Ghosh, M. Saketh, Niloy Ganguly, Bivas Mitra and Sandip Chakraborty

ComfRide:A Smartphone based System for Comfortable Public Transport Recommendation

Rohit Verma, Surjya Ghosh, M. Saketh, Niloy Ganguly, Bivas Mitra and Sandip Chakraborty Conferences

Passenger comfort is a major factor influencing a commuter’s decision to avail public transport. Existing studies suggest that factors like overcrowding, jerkiness, traffic congestion etc. correlate well to passenger’s (dis)comfort. An online survey conducted with more than 300 participants from 12 different countries reveals that different personalized and context dependent factors influence passenger comfort during a travel by public transport. Leveraging on these findings, we identify correlations between comfort level and these dynamic parameters, and implement a smartphone based application, ComfRide, which recommends the most comfortable route based on user’s preference honoring her travel time constraint. We use a ‘Dynamic Input/Output Automata’ based composition model to capture both the wide varieties of comfort choices from the commuters and the impact of environment on the comfort parameters. Evaluation of ComfRide, involving 50 participants over 28 routes in a state capital of India, reveals that recommended routes have on average 30% better comfort level than Google map recommended routes, when a commuter gives priority to specific comfort parameters of her choice.

Smart-phone based Spatio-temporal Sensing for Annotated Transit Map Generation

REDONDO BEACH - USA

Proceedings of the 25th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems(SIGSPATIAL 2017)

Conferences PDF PPT Rohit Verma, Surjya Ghosh, Niloy Ganguly, Bivas Mitra and Sandip Chakraborty

Smart-phone based Spatio-temporal Sensing for Annotated Transit Map Generation

Rohit Verma, Surjya Ghosh, Niloy Ganguly, Bivas Mitra and Sandip Chakraborty Conferences

City transit maps are one of the important resources for public navigation in today’s digital world. However, the availability of transit maps for many developing countries is very limited, primarily due to the various socio-economic factors that drive the private operated and partially regulated transport services. Public transports at these cities are marred with many factors such as uncoordinated waiting time at bus stoppages, crowding in the bus, sporadic road conditions etc., which also need to be annotated so that commuters can take informed decision. Interestingly, many of these factors are spatio-temporal in nature. In this paper, we develop CityMap, a system to automatically extract transit routes along with their eccentricities from spatio-temporal crowdsensed data collected via commuters’ smart-phones. We apply a learning based methodology coupled with a feature selection mechanism to filter out the necessary information from raw smart-phone sensor data with minimal user engagement and drain of battery power. A thorough evaluation of CityMap, conducted for more than two years over 11 different routes in 3 different cities in India, show that the system effectively annotates bus routes along with other route and road features with more than 90% of accuracy.

Unsupervised Annotated City Traffic Map Generation

SAN FRANCISCO - USA

Proceedings of the 24th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems(SIGSPATIAL 2016)
Project page Link

Conferences PDF POSTER Rohit Verma, Surjya Ghosh, Aviral Shrivastava, Niloy Ganguly, Bivas Mitra and Sandip Chakraborty

Unsupervised Annotated City Traffic Map Generation

Rohit Verma, Surjya Ghosh, Aviral Shrivastava, Niloy Ganguly, Bivas Mitra and Sandip Chakraborty Conferences

Public bus services in many cities in countries like India are controlled by private owners, hence, building up a database for all the bus routes is non-trivial. In this paper, we leverage smart-phone based sensing to crowdsource and populate the information repository for bus routes in a city. We have developed an intelligent data logging module for smart-phones and a server side processing mechanism to extract roads and bus routes information. From a 3 month long study involving more than 30 volunteers in 3 different cities in India, we found that the developed system, CrowdMap, can annotate bus routes with a mean error of 10m, while consuming 80% less energy compared to a continuous GPS based system.

Design Of Efficient Lightweight Strategies To Combat Dos Attack
In Delay Tolerant Network Routing

Wireless Networks, pp. 1-22
Springer US (2016)

Journal Paper PDF Sujoy Saha, Subrata Nandi, Rohit Verma, Satadal Sengupta, Kartikeya Singh, Vivek Sinha, Sajal K. Das

Design Of Efficient Lightweight Strategies To Combat Dos Attack In Delay Tolerant Network Routing

Sujoy Saha, Subrata Nandi, Rohit Verma, Satadal Sengupta, Kartikeya Singh, Vivek Sinha, Sajal K. Das Journal Paper

Delay tolerant networks (DTNs) are characterized by delay and intermittent connectivity. Satisfactory network functioning in a DTN relies heavily on co-ordination among participating nodes. However, in practice, such co-ordination cannot be taken for granted due to possible misbehaviour by relay nodes. Routing in a DTN is, therefore, vulnerable to various attacks, which adversely affect network performance. Several strategies have been proposed in the literature to alleviate such vulnerabilities—they vary widely in terms of throughput, detection time, overhead etc. One key challenge is to arrive at a tradeoff between detection time and overhead. We observe that the existing table-based reactive strategies to combat Denial-of-service (DoS) attacks in DTN suffer from two major drawbacks: high overhead and slow detection. In this paper, we propose three secure, light-weight and time-efficient routing algorithms for detecting DoS attacks (Blackhole and Grey-hole attacks) in the Spray & Focus routing protocol. The proposed algorithms are based on use of a small fraction of privileged (trusted) nodes. The first strategy, called TN, outperforms the existing table-based strategy with 20–30 % lesser detection time, 20–25 % higher malicious node detection and negligible overhead. The other two strategies, CTN_MI and CTN_RF explore the novel idea that trusted nodes are able to utilize each others’ information/experience using their long range connectivity as and when available. Simulations performed using an enhanced ONE simulator reveals that investing in enabling connectivity among trusted nodes (as in CTN_RF) can have significant performance benefits.

Margdarshak: A Mobile Data Analytics based Commute Time Estimator cum Route Recommender

SINGAPORE

Proceedings of the 3rd International on Workshop on Physical Analytics (WPA 2016)

Conferences PDF PPT Rohit Verma, Aviral Shrivastava, Samdip Chakraborty, Bivas Mitra

Margdarshak: A Mobile Data Analytics based Commute Time
Estimator cum Route Recommender

Rohit Verma, Aviral Shrivastava, Samdip Chakraborty, Bivas Mitra Conferences

Waiting at traffic signals and getting stuck in traffic congestion eats a lot of time for a commuter in most of the metro cities of the world. Although there exists a large pool of navigation applications, but all of them turn out to be ineffective for dynamically finding out the best route under uncertainty. In this work, we present Margdarshak, a navigation system which utilizes the impact of congestion and wait time at traffic signals for estimating the travel time over a route. We collected a month-long traffic data from different routes at five various cities in India for analyzing the problem in detail. The evaluations performed over the system show that Margdarshak gives a mean estimation error of ±1.5$ minutes, and performs significantly better under uncertainty, compared to other state of the art navigation systems like Google Maps, Here Maps and Waze.

UrbanEye: An outdoor localization system for public transport

SAN FRANCISCO - USA

Proceedings of The 35th Annual IEEE International Conference on Computer Communications (IEEE INFOCOM 2016)
Project page Link

Conferences PDF PPT Rohit Verma, Aviral Shrivastava, Bivas Mitra, Sujoy Saha, Niloy Ganguly, Subrata Nandi, Sandip Chakraborty

UrbanEye: An outdoor localization system for public transport

Rohit Verma, Aviral Shrivastava, Bivas Mitra, Sujoy Saha, Niloy Ganguly, Subrata Nandi, Sandip Chakraborty Conferences

Public transport in suburban cities (covers 80% of the urban landscape) of developing regions suffer from the lack of information in Google Transit, unpredictable travel times, chaotic schedules, absence of information board inside the vehicle. Consequently, passengers suffer from lack of information about the exact location where the bus is at present as well as the estimated time to be taken to reach the desired destination. We find that off-the-shelf deployment of existing (non-GPS) localization schemes exhibit high error due to sparsity of stable and structured outdoor landmarks (anchor points). Through rigorous experiments conducted over a month however, we realize that there are a certain class of volatile landmarks which may be useful in developing efficient localization scheme. Consequently, in this paper, we design a novel generalized energy-efficient outdoor localization scheme - UrbanEye, which efficiently combines the volatile and non-volatile landmarks using a specialized data structure, the probabilistic timed automata. UrbanEye uses speed-breakers, turns and stops as landmarks, estimates the travel time with a mean accuracy of ±2.5 mins and produces a mean localization accuracy of 50 m. Results from several runs taken in two cities, Durgapur and Kharagpur, reveal that UrbanEye provides more than 50% better localization accuracy compared to the existing system Dejavu, and consumes significantly less energy.

e-ONE: enhanced ONE for simulating challenged network scenarios

Journal of Networks 9.12, pp.3290-3304
Academy Publishers (2014)

Journal Paper PDF Sujoy Saha, Rohit Verma, Somir Saika, Partha Sarthi Paul, Subrata Nandi

Sujoy Saha, Rohit Verma, Somir Saika, Partha Sarthi Paul, Subrata Nandi Journal Paper

Delay Tolerant Network (DTN) empowers sparse mobile ad-hoc networks and other challenged network environments, such as interplanetary communication network or deep sea communication network, where traditional networking protocols either fail to work completely or do not work well. The Opportunistic Networking Environment (ONE) Simulator has gained considerable popularity as an efficient tool for validating and analysing DTN routing and application protocols. It provides options for creating different mobility models and routing strategies as per the users' requirements. Nowadays, challenged networks such as rural internet connection, social networks, post-disaster communication systems, etc. use DTN along with some hybrid infrastructure networks. Incorporating such real life network systems in ONE needs extensive modification of the same. In this paper, we present the enhanced ONE (e-ONE) simulator as an extension of ONE to facilitate simulation of challenged networks and describe the enhancements we have added to the ONE. As a case study, we consider a challenged network, which we call a latency aware 4-tier planned hybrid architecture designed for post-disaster management. We describe, in detail, how this enhanced version of the ONE simulator is useful in analysis and evaluation of the scenario considered.

Is It Worth Taking A Planned Approach To Design Ad
Hoc Infrastructure For Post Disaster Communication?

ISTANBUL - TURKEY

Proceedings of The Seventh ACM International Workshop on Challenged Networks (CHANTS 2012)

Conferences PDF Sujoy Saha, Vijay Kumar Shah, Rohit Verma, Ratna Mandal, Subrata Nandi

Is It Worth Taking A Planned Approach To Design Ad Hoc Infrastructure For Post Disaster Communication?

Sujoy Saha, Vijay Kumar Shah, Rohit Verma, Ratna Mandal, Subrata Nandi Conferences

After any natural disaster the availability of existing conventional communication infrastructure almost gets ruled out. After the devastation, to restore the communication system in ad hoc basis; ensuring almost 100% packet delivery within acceptable latency with optimal utilization of resources are prime design motives. Our work proposes a four tier planned hybrid architecture, which conforms the aforesaid motives yielding a desired performance in terms of delivery probability within least latency, for a given disaster hit area map with a suitable heuristic algorithm. Our study also reveals that there exists no deterministic polynomial time solution that can implement the desired design motives as well as the feasibility of our planned methodology. Compared to any brute force strategy, as per the simulation results, our approach shows 42% higher delivery probability and 49% lower latency.

SRSnF: A Strategy for Secured Routing in Spray and Focus Routing Protocol for DTN

CHENNAI - INDIA

Proceedings of the Second International Conference on Advances in Computing and Information Technology (ACITY 2012)

Conferences PDF Sujoy Saha, Rohit Verma, Satadal Sengupta, Vineet Mishra, Subrata Nandi

SRSnF: A Strategy for Secured Routing in Spray and Focus Routing Protocol for DTN

Sujoy Saha, Rohit Verma, Satadal Sengupta, Vineet Mishra, Subrata Nandi Conferences

This paper deals with the aspect of security in Delay Tolerant Networks (DTN). DTNs are characterized with decentralized control. Network performance and trustworthiness of transmitted information in DTNs depend upon the level of co-operation among participating nodes. As a result, DTNs are vulnerable towards untoward activities arising out of node selfishness as well as malicious intentions. In this paper, we limit our focus to the Black Hole Denial-of-Service attack. We develop a table-based strategy to record network history and use this information to detect discrepancies in the behavior of nodes, followed by elimination of those detected as malicious. We explain our detection mechanism considering Spray and Focus routing protocol as the representative routing scheme. The detection mechanism has been described in detail with examples pertaining to various case scenarios. Furthermore, we study the effect of variation of various parameters on detection efficiency and message transmission through simulation results.

ABOUT

RESUME

SYSTEMS RESEARCH SCIENTIST

Intel Labs, India

Doctor of Philosophy (PhD) - Computer Science and Engineering

INDIAN INSTITUTE OF TECHNOLOGY (2016 - 2020)

Bachelor of Technology (B.Tech) - Computer Science and Engineering

NATIONAL INSTITUTE OF TECHNOLOGY (2009 - 2013)

ADJUNCT FACULTY

INDIAN INSTITUTE OF INFORMATION TECHNOLOGY (2024-)

EXECUTIVE COMMITTEE MEMBER

IEEE Computer Society Bangalore Chapter (2023-)

RESEARCH ASSOCIATE

UNIVERSITY OF CAMBRIDGE (2020-2022)

JUNIOR RESEARCH FELLOW

INDIAN INSTITUTE OF TECHNOLOGY (2015)

SOFTWARE DEVELOPER

SCHNEIDER ELECTRIC INDIA PVT. LTD. (2013 - 2015)

INTERN

INDIAN INSTITUTE OF TECHNOLOGY (2012)

INTERN

INDIAN INSTITUTE OF TECHNOLOGY (2011)

POSTER CO-CHAIR

COMSNETS 2025

SysAI Workshop CO-CHAIR

SysAI Workshop, COMSNETS 2025

FINANCE CO-CHAIR

ICDCN 2025

TRACK CO-CHAIR

IEEE CONECCT 2024 - Computer Engineering & AI Track

POSTER CO-CHAIR

COMSNETS 2024

TRACK CO-CHAIR

IEEE CONECCT 2023 - Data Scale, Reliability, Security and Privacy Challenges in CPS Track

POSTER CO-CHAIR

COMSNETS 2023

POSTER CO-CHAIR

COMSNETS 2022

PUBLICITY CO-CHAIR

COMSNETS 2021

WEBMASTER

IMOBILE: The India Chapter of ACM SIGMOBILE

WEB CHAIR

MobiCom 2018

WEB CHAIR

COMSNETS 2018

Invited talk at the Research Scholar's Day RSD 2022 at National Institute of Technology Durgapur (Nov 2022).

Invited talk at the Workshop on "Last-mile" Challenges and Standardization Opportunities in Smart Infrastructure, COMSNETS 2022 (Jan 2022).

Technical talk session hosted by IEEE Communication Society- Delhi Chapter and IEEE Computer Society- Delhi Chapter (Aug 2021).

Supervisor - Operating Systems, University of Cambridge (Lent 2022)

Supervisor - Computer Networking, University of Cambridge (Lent 2022)

Supervisor - Computer Networking, University of Cambridge (Lent 2021)

TA - Computing Lab, Dept of CSE, IIT Kharagpur (Spring 2018 and 2019)

TA - Smartphone Computing, Dept of CSE, IIT Kharagpur (Autumn 2017)

TA - Computer Networks Lab, Dept of CSE, IIT Kharagpur (Autumn 2016)

TA - Programming and Data Structure, Dept of CSE, IIT Kharagpur (Spring 2016, Spring 2017, Autumn 2018)

COMSNETS Association Best Thesis Award - 2021

TCS PhD Fellowship - 2016

Google Student Travel Grant for IEEE PerCom 2019

LRN Travel Grant for IEEE PerCom 2019

Google Student Travel Grant for ACM SIGSPATIAL 2018

XRCI Travel Grant for IEEE INFOCOM 2016

ACM IARCS Travel Grant for IEEE INFOCOM 2016

PUBLICATIONS

Selective Graph Convolutional Network for Efficient Routing

Selective Graph Convolutional Network for Efficient Routing

On-the-Go Automated Break Recommendation for Stress Avoidance during Highway Driving

On-the-Go Automated Break Recommendation for Stress Avoidance during Highway Driving

Towards a Flexible Scale-out Framework for Efficient Visual Data Query Processing

Towards a Flexible Scale-out Framework for Efficient Visual Data Query Processing

Towards an Efficient Federated Learning Framework with Selective Aggregation

Bangalore, India

Towards an Efficient Federated Learning Framework with Selective Aggregation

FedNSE: Optimal Node Selection for Federated Learning with Non-IID Data

Bangalore, India

FedNSE: Optimal Node Selection for Federated Learning with Non-IID Data

Multi-Agent Packet Routing (MAPR): Co-Operative Packet Routing Algorithm with Multi-Agent Reinforcement Learning

Bangalore, India

Multi-Agent Packet Routing (MAPR): Co-Operative Packet Routing Algorithm with Multi-Agent Reinforcement Learning

DeepDish on a diet: Low-Latency, Energy-Eﬀicient Object-Detection and Tracking at the Edge

On-the-Go Automated Break Recommendation for Stress Avoidance
during Highway Driving

FedNSE: Optimal Node Selection for Federated Learning with
Non-IID Data

FedFM: Towards a Robust Federated Learning Approach For
Fault Mitigation at the Edge Nodes

Mining Spatio-temporal Data for Computing Driver Stress and
Observing Its Effects on Driving Behavior

Design Of Efficient Lightweight Strategies To Combat Dos Attack
In Delay Tolerant Network Routing

Margdarshak: A Mobile Data Analytics based Commute Time
Estimator cum Route Recommender

Is It Worth Taking A Planned Approach To Design Ad
Hoc Infrastructure For Post Disaster Communication?