“Wayve” is an uprising British A.I. enterprise that has developed a more efficient and effective way of autonomous training vehicles, and they believe that their mechanism is better than those methods which utilize and exploit equipment and 3D maps. In a video that they released, they announced and showed a renovated and remodeled Renault Twizy that can self-navigate.

The systems are equipped with a mechanism that is similar to a child where a form of reward and punishment system is observed. For example, when the machine displays good behavior, a reward will be given; conversely, if the machine exhibits bad behavior, a form of punishment will take place. The company believes that, through this kind of mechanism or approach, the machine or the vehicle will produce better results if contrasted to the other mechanisms.

Map Issues

A considerable part of these systems that are being developed nowadays is based on intricate and complicated 3D maps that are used for navigating. Companies are competing in the market to develop the most precise and accurate maps through the use of sensors and cameras that help them provide the self-navigating vehicles have the detailed features of a map.

Meanwhile, the AVs mentioned above (autonomous vehicles) still have to have their own set of cameras and sensors that are prerequisites for them to be able to function their navigation smoothly. However, they need to be updated periodically since changes occur every day.

Aside from that, the sectors or companies that are in charge of developing the maps often tend to overlook and disregard rural areas are they usually just concentrate and focus solely on congested and urban areas.

Minimization of the Technicalities

The new mechanism or system of learning could erase the need for things such as 3D maps. For instance, taking the Renault Twizy as an example only uses one camera in contrast to other vehicles that use up to eight cameras. With this, the minimum amount of camera is utilized to enable a feature that is very vital for a self-navigating vehicle – the GPU.

The GPU or the Graphics Processing Unit runs the reward and punishment mechanism or algorithm that enables the car to develop its way of learning how to navigate itself.

The car was first trained with a human driver for it to have a grasp on what are the good and bad driving behaviors; thus, triggering the car’s memory to record and utilize that data to develop its learning that was based under the supervision of the human driver. Due to the increased learning that was acquired by the car, Twizy exhibited excellent progress. This only proves that through this mechanism, vehicles can be able to ditch their dependence on 3D maps.