Pony.ai debuts its latest self-driving system: PonyAlpha. The release of PonyAlpha marks a significant milestone for China’s autonomous driving industry, representing the country’s first ever product-ready autonomous vehicle system.
PonyAlpha is a full-stack self-driving solution and has achieved levels of stability and performance capable of sustaining a consumer ride-hailing fleet, which has been expanding steadily in Guangzhou since February this year.
PonyAlpha features meaningful improvements across hardware and software from Pony.ai’s earlier generation prototype vehicles. In terms of hardware, the new system includes increased sensor coverage provided by additional LiDARs, radars, and cameras.
In order to support these additional sensors, PonyAlpha also features a more highly optimized hardware platform, resulting in a tightly integrated full-stack system. This sensor suite and configuration is a product of extensive testing across varied US and China environments.
Altogether, the full sensor suite can see distances of roughly 200 meters, enabling a wider variety of driving scenarios at even greater speeds.
The most notable change in PonyAlpha’s algorithms comes from its deep and adaptive sensor fusion module. Due to the multi-sensor approach (LiDAR, radar, camera), sensor fusion—the blending together of disparate data from multiple sensors—is an indispensable piece of autonomous driving technology, yet there is a natural tradeoff between amount of data collected and system simplicity.
Pony.ai’s sensor fusion technology is able to intelligently use the most reliable sensor data depending on different environmental or driving scenarios. Thus, Pony.ai vehicles have a comprehensive and accurate understanding of the surrounding environment while maintaining a highly effective on-vehicle process. The end result is not just a system that is highly accurate and efficient, but also one that delivers a smoother and more human-like driving experience.
However, building and operating an autonomous ride-hailing fleet takes more than just strong technology. Successful deployment and scale relies on an entire set of supporting infrastructure. In addition to rolling out PonyAlpha, the Pony.ai team has been working on several other critical fleet capabilities: the Pony.ai Vehicle Control Center and its in-vehicle interface PonyHI (Pony Human Interface).
The Pony.ai Vehicle Control Center enables centralized management, tracking, and dispatching of all vehicles in its fleet. For passengers, the launch of PonyHI gives them a direct window into the brain of the vehicle. Passengers are able to follow along the vehicle’s every move and decision as it identifies, classifies, and plans based on its surrounding environment. On screen, passengers can expect to see road agents such as vehicles, cyclists, or pedestrians as well as transportation infrastructure such as road signs or lane markings. As these objects appear and move on screen, passengers can also observe the “driver” take accurate and timely action, thus establishing transparency and trust between human and machine.