Motion planning

The motion planning team works on collision-free and comfortable trajectory planning for autonomous vehicles. We develop optimization-based approaches for generating safe trajectories in dynamic environments.

Our research includes sample-efficient collision risk minimization, trajectory optimization under stochastic dynamics, and robust planning methods that account for uncertainty in the behavior of other road users. We focus on methods that can handle real-time constraints while providing formal safety guarantees.

Arun Kumar Singh

Associate Professor of Collaborative Robotics, Motion Planning Team Lead

arun.singh@ut.ee

Research papers

Diffusion-FS: Multimodal Free-Space Prediction via Diffusion for Autonomous Driving MonoMPC: Monocular Vision Based Navigation with Learned Collision Model and Risk-Aware Model Predictive Control MMD-OPT: Maximum Mean Discrepancy-Based Sample Efficient Collision Risk Minimization for Autonomous Driving Trajectory Optimization Under Stochastic Dynamics Leveraging Maximum Mean Discrepancy LeGo-Drive: Language-enhanced Goal-oriented Closed-Loop End-to-End Autonomous Driving Talk2BEV: Language-enhanced Bird's-eye View Maps for Autonomous Driving Learning Sampling Distribution and Safety Filter for Autonomous Driving with VQ-VAE and Differentiable Optimization Hilbert Space Embedding-Based Trajectory Optimization for Multi-Modal Uncertain Obstacle Trajectory Prediction End-to-End Learning of Behavioural Inputs for Autonomous Driving in Dense Traffic UAP-BEV: Uncertainty Aware Planning Using Bird's Eye View Generated from Surround Monocular Images Bi-Level Optimization Augmented with Conditional Variational Autoencoder for Autonomous Driving in Dense Traffic Drift Reduced Navigation with Deep Explainable Features Multi-Modal Model Predictive Control through Batch Non-Holonomic Trajectory Optimization: Application to Highway Driving GPU Accelerated Convex Approximations for Fast Multi-Agent Trajectory Optimization Bi-Convex Approximation of Non-Holonomic Trajectory Optimization A Novel Trajectory Optimization for Affine Systems: Beyond Convex-Concave Procedure Model Predictive Control for Autonomous Driving considering Actuator Dynamics

Theses

Sample-efficient risk-aware trajectory optimization leveraging maximum mean discrepancy Learned Control Barrier Function for Reactive Navigation in Unknown Environments Risk-Aware Planning on Point Clouds Model-based Planning Using GPU-accelerated Simulator as a World Model Socially Aware Planning for Indoor Navigation Towards Reliable Real-time Trajectory Optimization Validation of NoMaD as a Global Planner for Mobile Robots Visibility aware navigation Sampling-based Bi-level Optimization aided by Behaviour Cloning for Autonomous Driving