Now Pony.ai released the third-generation self-driving truck software and hardware integration system. The system solution is designed for the needs of mainline logistics business. It has been applied to the first self-driving heavy-duty truck product jointly created by Pony.ai and Sany Heavy-duty Truck. The first smart cards.
According to the official introduction, Pony.ai has carried out all-round innovations in the software algorithms, sensor solutions, modeling design, and vehicle-by-wire chassis of the autonomous driving system, focusing on improving the stability and reliability of the heavy-duty truck products of autonomous driving vehicles under complex working conditions. Compatibility, to meet the subsequent mass production needs, in order to maximize the benefits of trunk logistics.
The next-generation system fully considers the needs of the autonomous driving trunk logistics business from the definition, development and mass production. Products with large-scale application capabilities of self-driving trucks.
The sensor solution of the third-generation autonomous truck system follows the multi-sensor deep fusion technology. The total number of sensors is 20, mainly using vehicle-grade sensor devices. Breakthroughs have been made in areas such as blind spots and long-tail scene processing.
Under the business requirements of high-speed driving and high load, the sensor solution consists of 1 ultra-long-range camera, 2 long-range cameras, 1 mass-produced vehicle-grade long-range lidar, and 1 long-range millimeter-wave radar. The ultra-long-distance perception module can cover the detection of obstacles from 200 meters to 1000 meters forward. Calculated at the truck’s 90 km/h driving speed, the ultra-long-distance perception module allows the autonomous truck to accurately identify the faulty vehicle ahead 30 seconds in advance, so as to avoid obstacles calmly.
Autonomous truck sensor solutions and their coverage
At the same time, in order to ensure the safety recall of obstacles within 200 meters around the body, especially the recognition of fallen cones, road potholes or thrown objects, the solution includes a 360° panoramic perception module, which consists of two 360° lasers. It consists of radar, 3 blind-filling lidars, 3 cameras, and 2 millimeter-wave radars, which are distributed on the front, rear and sides of the vehicle, solving the industry pain point of the near blind spot of the truck.
This system solution is also an important technical achievement since the cooperation between Pony.ai and Sany Heavy Truck was announced. Based on the autonomous driving prototype vehicle that the two parties have cooperated and verified in advance, Sany Heavy Truck is developing a new wire-controlled chassis for the first autonomous driving heavy truck product, fully considering the needs of autonomous driving, with full functions, low latency, high precision and strong robustness Features.
In terms of hardware design, the new generation of sensor kits adopts an exclusive integrated arc design (Sensor Arc), which improves the integration of key components and reduces the difficulty of truck installation and maintenance. Similar to the passenger car system solution released this year, Pony.ai has also added a Sensor Cleaning System to the truck system, and integrated aerodynamic principles into the industrial design of the sensor contact surface to effectively reduce dirt retention. It reduces the influence of rain, fog, sand, snow, and sludge that are common in trunk transportation on the sensitivity of the sensor, and can be applied to various working conditions.
In terms of software development, the new-generation system solution adopts a new sensor solution and algorithm design, so as to be flexibly compatible with various types of freight hanging boxes; in addition, the algorithm system can precisely adjust the horizontal and vertical control sensitivity of the vehicle according to the type of cargo and transportation requirements to ensure the protection of glass and ceramic tiles. Safe and smooth transportation of fragile items.
