Zhuoyu Company Adapts Autonomous Driving Model for Heavy Trucks; PDS 4 System Now in Mass Production
Read more
Pandaily
pandaily.com

Zhuoyu Company Adapts Autonomous Driving Model for Heavy Trucks; PDS 4 System Now in Mass Production

Zhuoyu Technology, formerly known as DJI Automotive, unveiled a commercial vehicle driver assistance solution at an event held in Zhuhai on October 29th. This solution is presented in two versions.

The PDS 4 version operates based on the company's driving model version 4.0 and later, providing Navigation On Autopilot (NOA) functionality for heavy trucks used on long-haul logistics routes. This version has already entered serial production.

The company plans to develop PDS 5, which will run on the next generation of Zhuoyu's native multimodal base model and will add urban NOA functionality for buses, light trucks, and vans.

Instead of creating a separate system specifically for trucks, Zhuoyu adapted its model developed for passenger cars. Company management described this approach as a unified foundation, scenario improvement, and engineering adaptation.

The company claims that its passenger car system has accumulated over 3 billion kilometers of driving data. Furthermore, approximately 400 million frames of data collected during highway travel were trained for the truck version.

Trucks present specific challenges: a fully loaded heavy truck can weigh around 40 tons, has large blind spots, and requires significantly more stopping time. Zhuoyu notes that the truck's braking chain has a delay approaching one second. The lidar system, installed inside the cabin rather than externally, reduces exposure to rain, snow, dirt, and stones. The company states that the system can detect obstacles on highways up to 500 meters away, including small objects with low reflectivity, such as tires and cardboard boxes.

The system also assesses trailer position, the angle between the tractor and trailer, and current load, coordinating the operation of the engine, retarder, auxiliary, and service brakes during ascents and turns.

Zhuoyu has additionally integrated a fuel-saving algorithm trained using reinforcement learning. This algorithm determines when to release the accelerator, apply power, or use the retarder based on gradient, curvature, powertrain characteristics, and surrounding traffic. Over two months of operation on more than 200 production trucks, covering over 20 million kilometers, the company reports fuel savings ranging from 4.5 to 7.5 percent.

The first serial vehicle is the Shacman X5000 from Shaanxi Automobile Group, production of which began in June. Several hundred such vehicles are currently being used on long-haul routes by the courier company Yunda. Zhuoyu is also collaborating with XCMG, Sinotruk, FAW Jiefang, Foton, and Dongfeng Commercial Vehicle, with several models expected to be released by the end of the year. The Yutong Bus will utilize PDS 5 on an urban bus.

It is important to note that this system remains a driver assistance system requiring constant monitoring, not fully autonomous control.

Popular