Curriculum Vitae

Ph.D. in Electrical and Computer Engineering

GPA: 3.85

B.Sc. (Hons.) in Electronic and Telecommunication Engineering

GPA: 3.97

Graduate Research Assistant

Project: Motion Planning and Perception methods for Robot Navigation

• Developed an algorithm for adaptive gait changes in quadruped robots that leverages cross-attention to dynamically fuse visual and time-series data, optimizing gait parameters for diverse terrain conditions.
• Contributed to the development of autonomous robot navigation algorithms that interprets human commands using vision-language models to enhance navigation and behavioral compliance in outdoor environments.
• Contributed to the development of an autonomous navigation algorithm that integrates vision-language models with proprioceptive sensing to assess terrain properties, improving navigational success rates in complex outdoor environments.

Technologies: Transformers, LLM, VLM, Cross-Attention, Representation Learning, Reinforcement Learning, ROS2, ROS, Pytorch, wandb

Undergraduate Researcher [Graduation Thesis]

Project: Behavioural and Local Planning and Maneuvering for Self-Driving

• Conducted a comprehensive literature review on Decision Making, Path Planning, and Control in Self-Driving.
• Developed a Finite State Machine to handle structured and unstructured environments.
• Implemented Path Planning using Optimization of Cubic Spirals and developed a Kinematic Model Predictive Controller for Lateral Control and a PID Controller for Longitudinal Control.
• Implemented parallel and perpendicular parking using Optimization Based Collision Avoidance.

Technologies: ROS, CasADi, CARLA, Tensorflow, Keras, MPC, PID, Cubic Spiral Path planning, A* algorithm, CNN, FSM, LiDAR, ZED2 Depth Camera, Jetson AGX Xavier

Undergraduate Researcher

Project: Anomaly Detection using Deep Reconstruction and Forecasting for Autonomous Systems

• Created real-time Unsupervised Learning Algorithms integrated with ROS to detect anomalous scenarios in UAV systems.
• Developed a CGAN optimized CNN-LSTM Autoencoder based Forecaster to detect anomalies in time-series frontal camera data.
• Developed an LSTM Forecaster and an LSTM Autoencoder Reconstructor to identify Instance and Transitional anomalies in time-series IMU data.

Achievement: Secured the 2nd position globally in the IEEE Signal Processing Cup 2020.

Technologies: ROS, Keras, Tensorflow, LSTM, GAN, CNN, AWS, Linux

Software Engineer

• Increased MFA capabilities of Asgardeo (CIAM IDaaS solution), by on-boarding the Duo Authenticator and sign-in with Microsoft, which increased the marktability of the product.
• Increased functionality and robustness of the Identity Server, an on-premise IAM solution, by separating audiences for Access Token and ID Token in OAuth2.0 flows.
• Contributed to the Architectural design and solutioning of a Consent Management System (Daisy) which will be deployed in the future.
• Ensured continuous functionality of the Identity Server 5.x/6.x through patches, and Asgardeo through product maintenance, which led to an increase in customer satisfaction.

Trainee Associate Electronics Engineer

• Conducted inference capability benchmarking on an Android POS device’s GPU/CPU, providing crucial insights into its performance and thereby offering the company valuable insights on the device’s suitability for Federated Learning initiatives. Furthermore, a PoC Federated Learning solution was implemented on the Android device.
• Researched on deep-learning based segmentation techniques and developed a real-time hair segmentation solution. This was implemented an Nvidia Jetson TX2.
• ncreased the efficiency of a Pose Estimation DNN which used VGG-16 as the feature extractor, by replacing VGG-16 with MobileNet, allowing inference in Android devices developed by Zone24x7 and thus lowering costs for additional GPUs.

George Corcoran Memorial Award for a Graduate Student

University of Maryland

Dean’s Fellowship

University of Maryland

2nd Place - IEEE Signal Processing Cup

ICASSP

Dean’s Honors List

University of Moratuwa

Mahapola Higher Education Merit Scholarship

Government of Sri Lanka

CROSS-GAiT: Cross-Attention-Based Multimodal Representation Fusion for Parametric Gait Adaptation in Complex Terrains

ICRA

Gershom Seneviratne, Kasun Weerakoon, Mohamed Elnoor, Vignesh Rajgopal, Harshavarthan Varatharajan, Mohamed Khalid M Jaffar, Jason Pusey, and Dinesh Manocha

BehAV: Behavioral Rule Guided Autonomy Using VLMs for Robot Navigation in Outdoor Scenes

ICRA

Kasun Weerakoon, Mohamed Elnoor, Gershom Seneviratne<>, Vignesh Rajagopal, Senthil Hariharan Arul, Jing Liang, Mohamed Khalid M Jaffar, Dinesh Manocha

VLM-GroNav: Robot Navigation Using Physically Grounded Vision-Language Models in Outdoor Environments

ICRA

Mohamed Elnoor, Kasun Weerakoon, Gershom Seneviratne, Ruiqi Xian, Tianrui Guan, Mohamed Khalid M Jaffar, Vignesh Rajagopal, and Dinesh Manocha

Anomaly Detection using Deep Reconstruction and Forecasting for Autonomous Systems

Nadarasar Bahavan, Navaratnarajah Suman, Sulhi Cader, Ruwinda Ranganayake, Damitha Seneviratne, Vinu Maddumage, Gershom Seneviratne, Yasinha Supun, Isuru Wijesiri, Suchitha Dehigaspitiya, Dumindu Tissera, Chamira Edussooriya

Behavioural and Local Planning and Maneuvering for Self-Driving

G. Seneviratne, V. Maddumage, Y. Madhushanike, R. Ranganayake, S. Herath