A synthetic LIDAR environment for self-driving cars.

There are an increasing number of sensors on modern vehicles for safety and efficient driving (ADAS) - these sensors and functions make it difficult to develop, verify, and test a vehicle. For example, to test a vehicle on a chassis dynamometers (CD) with a blower at the front of it, sensors/functions for collision avoidance need to be disabled. Currently it is possible to do this simply by turning off switches but in the future, this will not be possible for fully integrated autonomous vehicles. LIDAR is an indispensable sensor technology for autonomous vehicles and so a "cheating" device for use with a vehicle's LIDAR sensor will be needed for testing of such vehicles on a CD in a laboratory in the future.
LiDAR monitors shapes around vehicle by scanning a pulsed light source and measuring the time of arrival of reflected light on a detector, whereupon the distance to objects is calculated from the time-of-flight. It is desirable to
develop a system that can cheat a LiDAR system by absorbing the emitted light and reemitting a 'false' reflected light pulse. The device must be capable of matching the LiDAR field of view, responding fast enough (ns timescales) to cope with the operating rate of the LiDAR and would be expected to contain very similar hardware found inside a LiDAR, i.e. pulsed source, scanner/spatial light modulator, and detector.

I will work closely with Miles Padgett and Daniele Faccio at Glasgow University to evaluate the requirements for such a device to cheat a commercial LiDAR system. We will build and test a proof-of-principle system in laboratory
conditions.

Successful completion of this prototype could be further evaluated at the MIRA test facility to investigate the operation of the device on a vehicle on a chassis dynamometer.