Rakshit Soni

DIVE, Driverless and Intelligent Vehicles. Rakshit Soni · autonomous vehicles & human interaction

scroll to dive in

PhD researcher · Automotive engineer

Building autonomous vehicles that get along with people.

I'm Rakshit Soni. I work on autonomous vehicles and human interaction, and I build on real vehicle software stacks in ROS2.

See the projects Read insights

Before · United Kingdom

Learning to research

Master's by Research in Computer Science at the University of Lincoln, in the L-CAS robotics group. I studied game-theoretic interaction between pedestrians and autonomous vehicles, and helped build the open-source OpenPodCar_V2.

where people met robots →

Now · France

Where I work today

CIFRE PhD with Stellantis, the CESI LINEACT laboratory and ENSAM, working on autonomous vehicles and human interaction.

more to share once it's published

Who am I

Automotive engineer turned automated-driving researcher.

I'm a CIFRE PhD researcher working on autonomous vehicles and how they interact with people, alongside industry and academic partners in France.

Before that, I was a Master's by Research candidate in the L-CAS group at the University of Lincoln, UK, studying game-theoretic interaction between pedestrians and automated vehicles.

  1. Now
    CIFRE PhD Stellantis · CESI LINEACT · ENSAM, France
  2. Before
    MSc by Research, Computer Science University of Lincoln, L-CAS, UK
  3. Start
    BEng, Automotive Engineering
More about me →

Selected projects things I've built ↓

From research questions to running vehicles

All projects →

University of Lincoln · L-CASOpen source

OpenPodCar_V2, an open-source autonomous vehicle platform

A low-cost autonomous vehicle built from a commercial mobility scooter, with open hardware and a full ROS2 software stack for human–AV interaction research.

ROS2NAV2RTAB-MapGazebo

OpenPodCar_V2 perception stackField tested

Real-time 3D pedestrian detection and tracking in ROS2

A ROS2 wrapper that turns YOLOv8 detections from an RGB-D camera into 3D pedestrian positions and velocities the vehicle can plan around.

YOLOv8ROS2Intel RealSense D435Python

University of Lincoln · EU INTERACT projectResearch

Game-theoretic pedestrian–AV crossing decisions

Modelling a pedestrian and an automated vehicle at a crossing as a game, so the vehicle can reason about probable outcomes before it acts.

Game theoryNashpyPythonROS2

Latest insights

Technical write-ups, feature analysis and white papers

All insights →