Volvo's technology, the result of a research project called 'Non-Hit Car and Truck', is currently in the test phase. The main component of the technology is a sensor data fusion platform that receives and processes the sensory input from cameras, radars and other sensors located on all sides of the truck. The platform creates a 360-degree scan of its surroundings every 25 milliseconds. 360° surround view applications have been available before Volvo's development, but these typically are just stitching together the images of four cameras located at the corners of the car, generating a virtual bird's view. Such simple systems are used for situations when the car is moving along very slowly, for instance during parking.
Volvo's system is more sophisticated in that it uses sensor data from multiple types of sensors and processes them at higher speed, enabling the platform to be useful also in moving traffic. The algorithms interpret the data input, detect and analyse risk situations and generate alternative trajectories for the vehicle. The combination of different sensory systems enables Volvo's platform to distinguish and identify different road users including pedestrians, cyclists, motorcyclists and other vehicles. It can predict traffic scenarios up to five seconds ahead, depending on the speed of the objects, and map the best plan of action. While it in a certain sense acts as a co-driver, it also enables the driver to see everything around the vehicles.
In the case the system detects the hazard of a collision, it alerts the driver. In the current stage, the system is in the test phase; it is not yet ready to applied in a commercial safety system due to the specific challenges to overcome when used emergency assists in heavy vehicles. "Trucks are a different type of vehicle and do not act the same ways as cars in traffic", said Mansour Keshavarz, systems engineer at Volvo Trucks. Factors like loading vary from vehicle to vehicle,