Hardware accelerated collision checking for robotics

The ability to quickly check if robot poses are not in collision with environment geometry is critical for robot applications that rely on forward kinematics, inverse kinematics, and path planning. It has been shown that during path planning 99% of the time is spent performing collision checks between the robot & its environment. Improving collision checking speed can have a dramatic impact on the feasibility of planning in real-time and the quality of the plans produced. The objective of this project is to prototype a collision checking library that can utilize the faster vector operations of a specialized compute like GPU.

Faculty Supervisor:

Nandita Vijaykumar

Student:

Partner:

Kindred AI

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of Toronto

Program: