Robot that sees through (cardboard) walls

I made a Python version of the library for communicating with grove’s Doppler radar: GitHub - tooploox/grove-doppler-radar: Python library for communication with grove doppler radar based on the BGT24LTR11 and wanted to try it on my car-like robot.

I decided to implement a simple driver assistance system that would stop a car (or in reality, just warn a driver) if it senses a fast-approaching object on a collision path, even if the object is fully occluded by a wall or other obstacle.

My radar’s signal can’t pass through concrete walls but a wooden desk or cardboard doesn’t seem to be a problem. And, the technology for sensing objects through walls does exist, it would be convenient to have it in a car! The ability to measure the speed of other cars would be a bonus.


Demo code: GitHub - aslucki/radar_ADAS
Car’s model code (it runs on Jetson Nano): GitHub - tooploox/autonomous_car_model: Physical and simulated model of a car for developing and testing autonomous driving and driving assistance solutions.


This is a nice idea. Would love to see it with an even more powerful ground penetrating radar (probably not practical due to cost, plus it wouldn’t fit on a tiny model). Or with a display augmented reality camera to show the driver visually what the trigger was.

Thanks! You’re right, AR would be great, or at last head-up display. I was thinking that an even less powerful radar or set of radars covering the front and front sides of a car could be quite useful as well. Knowing the speed and distance of other moving objects would make many maneuvers easier.

I’ve seen some usable automotive radars for ~$600, it could be feasible for carmakers to install radars in their cars.