Online shopping service tests unmanned delivery vehicle

REMOTELY CONTROLLED, the self-driving vehicle covers a 4-kilometre route, from a distribution centre to a Collect&Go pick-up point.

The Clevon 1 vehicle works on the basis of vision technology, machine learning and sensors. In a suitable controlled environment, the vehicle, which can travel up to around 50km/h (but is agreed to keep it under 25 km/h for the initial test phase), can be operated in the autopilot mode. During this, a large number of general driving tasks will be performed by the vehicle autonomously with the teleoperator supervising the vehicle at all times. 

The teleoperator can also take over the control of the vehicle at any given time to ensure the highest level of safety. To observe its surroundings, it uses three front cameras, two side cameras and one rear camera. Beside the cameras, the vehicle is also equipped with short- and long-range radars, which can measure distances and recognise and identify "obstacles" such as cars, cyclists and pedestrians, among other things. In this way, it has a clear field of view of its surroundings and can predict and thus avoid possible collisions in traffic. Clevon uses multiple deep neural networks, fusing together camera and radar information which allows the vehicles to detect and identify the dynamic environment.