Collaborative Perception and Planning Models for Smart Infrastructure and CAVs

Status: In Progress

Lead Researcher

Gregory Stevens
President, Causality LLC; Research Director, Mcity

Project Team

Kate Skinner
Ram Vasudevan

Project Abstract

Demonstrate the unique incremental benefits of integrating infrastructure perception into vehicle decision-making in Depot Autonomy. Within the constrained ODD of a depot, allow an L2 ADAS vehicle to operate as an L4 AV by adding infrastructure perception and routing.

Project Outcome

  • Create collaboration models for Level2 ADAS vehicles with smart infrastructure to allow Level4 autonomy within a delivery van loading depot (eg- Fedex, DHL, UPS, …)
  • Demonstrate a working, virtual prototype of Depot Autonomy.
  • Leverage implicit neural representations for multi-sensor fusion of high fidelity infrastructure-mounted LiDAR and low cost vehicle-mounted cameras to enable collaborative perception for commercial environments with dynamic objects.
  • Implement computationally-efficient, real-time motion planning with provable safety.

BUDGET YEAR: 2024-01-01