While companies like Waymo and Uber use an expensive radar-like system called LIDAR (Light Detection and Ranging) in their self-driving car programs, AImotive claims to achieve the same using regular cameras combined with artificial intelligence. The company claims that this brings down the cost of converting a regular car into a driverless one to around $6,000 as opposed to $70,000-$100,000.
“The whole traffic system is based on the visual system. Drivers don’t have bat ears and sonars, you just look around and drive,” said Laszlo Kishonti, CEO and Founder, AImotive.
According to Kishonti, LIDAR has become the industry standard system for driverless vehicles because of its success in DARPA’s driverless car competition 10 years ago.
Howerver, the AImotive car is installed with four fish-eye cameras on each side of their vehicle, as well as two dual stereo cameras on the front and back of the vehicle. The boot of the car contains a powerful PC that is designed to create a three-dimensional model of the car’s surroundings by stitching together the feeds from the cameras in real time.
Decisions about the surroundings are made using the AI technology, depending on the kind of surrounding the car is operating in. The car’s “location engine” uses GPS technology to situate the vehicle on a map, ensuring precise self-localization and navigation, while also keeping the in-car memory requirements extremely low.
An AI-driven “motion engine” tracks and predicts the motion patterns of surrounding objects, and combines navigation output from the Location Engine with the predicted state of the surroundings to decide the right trajectory for the vehicle. It controls the cars movement on the road, and feeds into a “control engine” that determines steering, acceleration, and braking. Motion Engine can also recognize emergency situations and act accordingly with a secondary trajectory plan that claims to ensure maximum safety for both drivers and pedestrians.
Although AImotive is yet to receive its California driverless car license, it has tested the modified Prius on roads and car parks in Budapest, where it is headquartered. The company is currently testing its technology with manufacturers like Volvo to provide driverless systems for cars and trucks. It plans to begin testing its system in urban settings in early 2018.
