//
In recent years, the technology and management model of the mining industry has gradually changed and upgraded with the popularity of the metaverse and the development of autonomous driving technology, and then derived the concept of meta mines, in which audio and video technology has also become a key link. LiveVideoStackCon 2023 Shanghai Station invited Ren Siliang from Xidi Zhijia to share with you the history of Xidi Zhijia's meta-mine construction and the application of audio and video technology in the meta-mine system.
Text/Ren Siliang
Organize/LiveVideoStack
Hello everyone, I am Ren Siliang from Xidi Zhijia. I am mainly engaged in the research and development of autonomous driving technology and the implementation of products. You may still be relatively unfamiliar with Xidi Zhijia, so let's give a brief introduction to our company.
-01-
Company Introduction
Xidi Zhijia was established in 2017, also known as Changsha Intelligent Driving Research Institute Co., Ltd. It is a unicorn company founded by Professor Li Zexiang, the former chairman of DJI Innovations. Enterprises such as Cloud Whale Technology, Zhenghao EcoFlow and Zuvi Yuanli It is also an incubation company under Professor Li's entrepreneurial system.
Xidi Zhijia currently has more than 500 employees, of which the R&D team accounts for nearly 70%, and the team's doctors and masters account for more than 40%. It is a technology company with technology research and development as its core driving force.
Our main business revolves around the autonomous driving sector, covering legally permitted unmanned driving technologies in closed environments such as park logistics vehicles, sweepers, port trucks, and mining trucks, as well as intelligent V2V and V2X car networking technologies that ensure the safety of unmanned vehicles .
-02-
Yuan Mine Introduction
In the past two years, the increasing demand for autonomous driving of mining trucks has given birth to the cloud mining concept of Xidi Zhijia. The birth of unmanned mining trucks starts with a social event. You may know something about the collapse of the Xinjing Coal Mine in Alxa Left Banner, Inner Mongolia in February 2023. So far, 47 workers have been buried tens of meters deep.
While the traditional mining industry occupies an important position in the national economy, it also faces challenges such as high risks, high energy consumption, aging practitioners, and low work efficiency.
On the basis of fully researching the current pain points of the mining industry, Xidi Zhijia launched the self-developed unmanned pure electric mining truck. It combines intelligent dispatching system, V2X vehicle-road coordination system and remote driving system to help the construction of smart mines. The main components of the Dizhijiayuan mine system. The picture above shows the 14 real vehicles we officially delivered to the Taiwan Cement Mining Area in Jurong City, Jiangsu Province.
From September 2022, the Jurong mining area will realize normal production operations without safety personnel.
Our unmanned mining truck is the only automatic driving system in the country certified by the National Institute of Metrology to exceed human efficiency. Calculated based on the loading capacity of a single vehicle, its automatic driving efficiency is 1.04 times that of manual labor.
Considering vehicle maintenance, energy, labor and other factors, the operating cost of manned diesel mining trucks is 8.6 times that of autonomous driving.
The picture above shows our "vehicle-road-cloud" overall solution for Yuan Mine. The entire process from excavator loading, mineral transportation, crushing station unloading to parking lot charging has been fully unmanned. The implementation of this solution enables traditional mines with 20 to 30 employees to reduce the number of operators to 2 to 3, and the daily management of unmanned vehicles can all be completed in the dispatch center.
The picture above shows the on-site demonstration effect of dynamic loading of unmanned mining trucks. The vehicle can dynamically adjust the loading and parking position with the continuous changes of the mining operation surface of the mine, which is also a key link of the Yuan mine system.
During the transportation process, the uniform rolling function of the unmanned vehicle can avoid deep ruts caused by repeated rolling of the road surface, which is conducive to maintaining the smoothness of the road and reducing maintenance costs.
Compared with traditional mines, Yuankuang mine unmanned mining trucks support all-weather automatic driving operation, which further improves work efficiency.
In terms of the application of audio and video technology, first of all, in order to deal with the few scenarios that may be difficult to deal with by automatic driving, we have implemented a remote driving solution with the help of Tencent Cloud's low-latency transmission protocol, which can make the command delay from the cockpit to the vehicle less than 100 milliseconds , to ensure that the cabin end operator basically does not feel the delay.
The solution also includes L2-level driving functions such as one-key slope repair, dynamic homing, and anti-collision. The display at the cabin end applies our self-developed naked-eye 3D algorithm, which can realistically display the obstacle environment in front of the vehicle and help improve work efficiency.
In addition to mining trucks, this solution can also be applied to excavators, drilling rigs, transfer machines or explosive placement robots to further ensure production safety and realize unmanned operations.
Next, we will show several practical application scenarios of remote driving. As shown in the figure above, if the vehicle encounters an unavoidable obstacle during automatic driving, it can switch to the remote driving mode for processing.
The picture above shows the practical application of unmanned driving technology in the 1,000-meter underground environment of Jinchuan No. 2 Mine in Gansu Province. The driver can grasp the vehicle's position in real time through the remote cockpit screen, ensuring safe production.
Combining V2X vehicle-road coordination and automatic driving technology, we have implemented the mixed production and transportation of 30 unmanned pure electric mining trucks and manned vehicles at the Daye Building Materials Mine in Fuling, Chongqing. Making full use of the perception capability of the V2X system can realize L3, L4, and even L5 automatic driving of unmanned vehicles in a closed environment.
The picture above shows the application of remote driving technology in the scene of sweeper and port truck.
Similar to the VR glasses in the Metaverse, we hope that the remote cockpit can also become the entrance to the Meta Mine, thereby helping to improve the working environment.
Through automatic driving + remote driving, we hope that young people can work remotely in a "comfortable and gamified" way, and solve problems such as high risks in traditional mining and difficulty in recruiting workers.
The second application of audio and video technology in the Yuan mine system is web live broadcast, which is mainly used for external publicity, customer explanations, unmanned mine monitoring and other scenarios. The picture above shows the demonstration effect of the unmanned monitoring scene. Through the monitoring screen, the operator can view the environment in front of the vehicle, driving track, speed, gear position, loading mode and transportation mileage in real time. The driving boundary represented by the yellow line in the left picture can be The on-site environment is dynamically detected with the help of body radar for real-time updates.
The third application of audio and video technology in the Yuan mine system is image recognition. Different from Tesla's use of image recognition for on-board decision-making, we use it more to increase the safety redundancy of vehicles. Through the perception fusion of various sensors such as millimeter-wave radar, laser radar, vision, and ultrasonic radar, the integration of several sensors Advantages Realize a 360-degree range of the body, and stable perception of obstacles such as pedestrians, vehicles, and stones 100 meters ahead.
As shown in the figure above, we complete obstacle recognition by fusing 2D visual images and 3D SLAM point cloud mapping, which effectively saves CPU resources.
The picture above shows the field effect of image recognition. It is not the main means of obstacle recognition due to the influence of severe weather factors such as dust and haze in mining areas. Image recognition and detection under severe weather such as smog is also a major technical challenge we are currently facing.
Our 360-degree intelligent surround view system can collect high-definition video images around the vehicle body in real time to form a three-dimensional panorama. At the same time, it can detect vulnerable traffic participants and give them graded warnings, provide driving trajectory guidance, and ensure driving safety.
The fourth application of audio and video technology in the Yuan mine system is visual positioning. It is similar to the fusion perception mentioned above. For scenes where traditional GPS+IMU solutions such as tunnels are difficult to accurately locate, the camera pre-scans the environmental features and establishes a SLAM point cloud map, which is realized according to the positional relationship between the vehicle and the surrounding feature points when driving. position. However, this solution is not suitable for scenes such as highways where the scene heights on both sides of the vehicle are similar. In actual work, we generally use a combination of GPS+IMU, RTK differential positioning and visual positioning.
-03-
Project cases
Finally, I would like to introduce a few project cases. The first is the metal mine of Jinchuan No. 2 Mine in Gansu Province. It is the largest nickel mine in China. The operating environment is located below 1,000 meters underground. The delivery of unmanned mining trucks has greatly improved the working environment of mining personnel. .
In August 2022, we delivered 4 unmanned pure electric mining trucks and supporting systems to Datong Jidong Cement Mine, and the mining area has entered normal production.
Our unmanned pure electric mining trucks are generally suitable for the transportation of cement mines and other sand and gravel aggregate mining areas. This is because the mining trucks can recover kinetic energy during the top-to-bottom transportation of sand and gravel aggregates, which is conducive to cost reduction synergy.
Chongqing Fuling Daye Building Materials Mine driverless mining truck project is currently the largest pure electric wide-body vehicle automatic driving project in China. In March 2023, 10 real vehicles have been delivered and the trial run has been completed.
It took only 3 months to complete the deployment of 14 unmanned pure electric mining trucks in Taiwan's cement mining area in Jurong, Jiangsu, and has achieved 7x24 unmanned operation.
The ultimate goal of our Yuan mine is to realize a mine that can be mined remotely, and our vision is to make the world have no difficult mines. That's all for my sharing today, thank you all.
▲Scan the QR code in the picture or click " Read the original text " ▲
Direct access to LiveVideoStackCon 2023 Shenzhen Station 20% off ticket purchase channel