Remote Control – or Remote Operation - is the operation of a vehicle (car, robot, drone) by a human who is located somewhere else. The human operator, or remote driver, operates a steering wheel and gas/brake pedals before a screen(s) that shows the vehicle’s direct surroundings.
This Remote Control is essential for fleet managers to bring rental / shared cars from locations of low demand to locations of high demand (e.g. an airport). Furthermore, it is used for edge cases’ where autonomous cars are stuck and need a human to overcome the situation. Thirdly, it’s legally mandatory in some countries to have this option with an autonomous vehicle.
Key for this is of course low transmission times of the video from car/robot to control center as a delay between what the human driver sees and what the car is doing can be fatal. A more operational benefit is the significantly reduced costs for mobile network (3G/4G) usage.
The model provides quick and accurate region of interest detection and supports real-time integration of multiple sensor data together such as 3D point cloud, HD camera etc. The model can enable a 40ms end to end latency for up to 4-5 camera + 2 LIDAR streams.
Teraki’s Remote Control model can deliver a up to 40x reduction of sensor data that needs to be transferred to the cloud, thereby improving the latency for enabling real-time LIDAR/RADAR/camera based remote operation of the car and saving the bandwidth consumption cost.
The Platform enables customers to (re)train their own ROI/TOI models. With such tailor-made detectors, customers can refine the next iterations of data collection and with this precizer targetting of objects and events, this will result in updated AI-models with higher accuracy.
How to Implement with Teraki:
ROI and TOI models are available on the Platform to be used: https://www.teraki.com/platform