Autonomous cars are finally on their way. Experts predict that within a span of five years, we will see self-driving vehicles become mainstream on the roads. Automotive players around the world are leveraging on this opportunity with their autonomous driving solutions. One such innovative solution is DragonFly autonomous vehicle, created by PerceptIn.
Dr Shaoshan Liu, founder and chairman, PerceptIn, created DragonFly to make low-cost autonomous vehicles available for use. His vision was to create self-driven vehicle that was affordable and could aid the elderly, physically-challenged or those who could not drive. Utilising his previous experience with automotive artificial intelligence (AI) programs and sensors, he went on to design DragonFly.
DragonFly is a two-seater autonomous vehicle pod powered by a vision-based, multi-sensor and fusion-based system. It uses computer vision, machine learning, AI and analytics to drive itself across a route.
The invention is intended for low-speed use cases, like in campuses, less busy roads or shorter routes. It can be used to transport two passengers or a limited amount of goods at a time.
A proprietary mobile application can be used to summon the vehicle. When called, the vehicle travels from its current location to the destination where the user is standing, utilising its inbuilt navigation feature.
Machine learning algorithms help the vehicle identify the fastest route to the destination. The vehicle comes with multiple proprietary sensory products created by PerceptIn, including DragonFly computer vision module on top, Global Navigation Satellite System (GNSS) modules next to the computer vision sensor, and sonar and 77GHz millimetre-wave radar on the front.
Inputs through each of these modules drive various functionalities of the vehicle, including object and proximity detection, intersection path planning, signal detection, obstacle avoidance and curvature planning, pedestrian detection and so on.
DragonFly has night-vision detection capability to drive in low-light conditions. In case of complex on-road situations, it shifts its control to the central control station remotely, to be driven manually by the operator.
DragonFly pod is deployment-ready. It is available for US$ 40,000. However, the team looks forward to a much-reduced price tag once demand picks up.
PerceptIn also delivers various modules and the necessary software development kit (SDK) individually, which can be used by automakers to create their own modular autonomous vehicles. The company claims that this reduces overall construction and maintenance costs, and substantially cuts down time-to-market.
PerceptIn is also working on autonomous exploration vehicles, suitable for harsh environments that are usually unsafe for human explorers. Advertisement vehicles are also part of their product portfolio. For technopreneurs looking to try their hands at autonomous vehicle design, PerceptIn team has composed a book and are developing courses with its IEEE community.