The solution uses an in-vehicle computer (LEC-2290H) with edge AI capabilities to pre-process data on the vehicle and make decisions in real-time. It also uses a rugged network appliance (ISD-O370) that supports 5G so that all data can be moved fast to the cloud for remote control and monitoring.
Mine operators face countless challenges, but one of the most significant ones is finding the right people to drive mining vehicles. Labor costs sharply increase, not to mention the risk of involving more personnel in transportation and road operations. Humans can get tired, drive slightly off-course, or make simple mistakes.
An obvious solution to reducing recruitment challenges and improving safety is introducing automation. Autonomous driving mining vehicles can be a highly effective solution as they not only reduce human error but also introduce innovation and enhance productivity. In addition, it is also easier to implement automation in “predictable” and “controlled” environments such as mines.
Although autonomous mining trucks operate in these “predictable safer environments,” they are still heavy-duty driverless machinery carrying tons of materials and working around other machinery and personnel on site— so safety is no exception. Operating these mining trucks with zero tolerance for error is therefore critical.
An autonomous driving mining vehicle should operate in real-time, with high precision, efficiency and security. So getting the right type of technology to these trucks is paramount. If mine operators want to leverage autonomous technology, they would need the suitable computing and networking infrastructure that can operate in challenging environments, especially when it comes to dust and humidity.
Autonomous Driving Mining Vehicle Solution.
Mining companies should leverage different technology components to integrate a high level of automation into mining operations. There should be a collaboration between vehicles, devices, roads, networks, and the cloud.
The below diagram shows how the autonomous driving mining vehicle solution integrates some of these components (Vehicle > Network > Cloud).
Autonomous mining trucks should be able to act in real-time and perform quick advanced decisions. They should also sense the environment with high precision and navigate through a mine without human interaction.
Equipment deployed on mining vehicles should therefore be high-performance, stable, and durable to withstand harsh operating conditions.
- In-vehicle data collection. Various devices, including sensors, Radar, LIDAR, and cameras deployed around the vehicle, can collect data from its surroundings. The system can also benefit from high-precision GPS to create a picture of a truck’s location and speed. In the vehicle, all the data should also be processed (compiled, deduplicated, or reduced) so that the vehicle can make immediate and smart decisions without depending on the network.
- In-vehicle computing. The mining vehicle can use an in-vehicle edge AI solution such as LEC-2290H to empower autonomous driving software. AI chipmaker Hailo created Hailo Falcon-H8, an AI-powered PCIe accelerator card. Falcon H8 is a low energy consumption PCIe with an exceptional performance of 156 Tera Operations Per Second (TOPS). This performance allows deep learning and model training to be deployed on vehicles. Lanner’s edge AI appliance (LEC-2290) provides the PCI interface for Falcon H-8 and the capacity to perform AI on edge. LEC-2290 is an Nvidia NGC-ready edge AI appliance with support for Intel® Core i7-8700T/i7-8700 CPU. LEC-2290 is also designed to operate on varying temperatures, relative humidity, and vibration.
Although the autonomous mining vehicle uses an in-vehicle edge AI solution, which could make a vehicle independent from an internet connection, reliable wireless network communication could also introduce benefits. For instance, low latency and high bandwidth connections such as 5G can help with remote support or emergency takeover. Plus, this wireless connection can also help to upload HQ video data to the cloud or high-end server. Network infrastructure deployed on open-pit mines should also be able to withstand harsh conditions.
- A rugged network appliance. Lanner’s ISD-O370 is an IP67 rugged 5G edge network appliance capable of withstanding the harsh and dusty conditions of mines. This appliance can be deployed on the roadside to provide wireless communications (5G or WiFi6). In addition, ISD-O370 can also provide edge AI capabilities to support the in-vehicle edge AI solution suggested above. Bringing intelligence as close as possible to where mining vehicles are working, such as roadside, nearby closet, or external data center, can make autonomous mining vehicles more reliable and safe.
Although the edge AI system processes and analyzes data on-site, the system could still use the cloud for collaborative and remote operations— among other benefits. Delivering pre-processed data to the cloud can help remote mining operators and managers. The cloud can provide services such as autonomous driving cloud service (SaaS), V2X (Vehicle to Everything), HD map cloud service, and more. Cloud-based software for vehicle control and monitoring can also help remote mine operators coordinate vehicle routes, queuing, scheduling, and more.
Autonomous driving mining vehicles can operate in remote (unmanned) sites with no on-site personnel and almost regardless of environmental conditions. One agent with access to the cloud-based control center and software could oversee the mine operations. This level of autonomy can help mining operators improve safety, productivity, and savings.
- Improve safety. Although humans could drive mining vehicles with a great degree of safety, autonomous mining vehicles can be configured to drive the same route many times without getting tired or losing focus. Eliminating human risk to error improves safety, not only for drivers but for other fellow workers and equipment. Having fewer humans on mining sites also reduces the exposure of mining workers to more hazardous conditions.
- Improve productivity with advanced functions. Introduce advanced capabilities such as environment identification and autonomous positioning. In addition, the system can also use Deep Learning (DL) to improve scheduling and optimize transportation efficiency. The autonomous mining vehicle can also perform operations such as loading/unloading, parking, and keeping a route— all independently from humans.
- Ongoing monitoring and predictive maintenance. The autonomous driving mining vehicle system can also use the collected data for real-time vehicle monitoring and predictive maintenance. These actions allow operators to save on fuel costs, spear parts, expensive overhauls, tire wear, and overall labor costs.
- Improve savings. Although there is a high initial cost of investment to implement autonomy in mining vehicles, the increased efficiency and productivity of the system will improve savings and increase revenue. Autonomous driving mining vehicles can be programmed to constantly operate the vehicle at optimal speed and use the best routes.
Next Steps …
For more information on another edge AI in-vehicle or network appliances or the autonomous driving mining vehicles solution, please contact Lanner’s sales representative.
- Edge AI Computing for Autonomous Driving System
- AI-powered Edge Computing Platforms Enable Autonomous Driving
Intelligent Edge Appliance With Falcon-H8 PCIe AI Accelerator Card
|CPU||Intel® Core™ i7-8700 (Codenamed Coffee Lake S)|
High-performance And Cost-efficiency PCIe AI Accelerator Card
IP67 Rugged 5G Edge Network Appliance
|CPU||Intel® Atom® C3000 (Denverton)|