Antonio, Simbarashe (2020) ST18 Autonomous enhancement strategies (detection of wall impacts and development of a self-cleaning laser scanner). [USQ Project]
|
Text (Project)
Antonio_S_P_Wen_Redacted.pdf Download (5MB) | Preview |
Abstract
Fatalities and major accidents continue to occur in mines. While technology has continued to play a pivotal role in assisting mining companies to safely and successfully conduct their operations, a high level of innovation is still required to attain greater technological advancements. The industry needs competent and skilled personnel to design, implement and rectify processes and systems within the mining industry to make autonomous systems safe and reliable.
Epiroc Australia (Epiroc) is in manufacturing front-end loaders that are highly optimised for underground mining environments to perform the load-haul-dump (LHD) function of the mining process. This family of machines is known as Scoop Trams (ST) and comes in various bucket capacities from seven metric tonnes up to 18. To complement the ST-family of LHD loaders, Epiroc has developed an autonomous system for the ST18 to enhance safety, workplace environment and productivity for clients in underground operations. The autonomous system's challenges in an underground mining operation system impact aspects that include the environment, network systems and dust and rock structure instability.
The autonomous system utilises the Kalman filter for position localisation using the odometer wheel sensor, articulation angle sensor, Inertial Measurement Unit (IMU) and two LIDAR scanners. The Kalman filter runs recursively to continually update the loader's position by optimally estimating the system variables (speed, attitude, gyroscopic data, acceleration and the laser scan data). If any of the sensors fails, the automation system will detect this and stop the machine to prevent any damage from occurring. This research paper focused on the reliability issues caused by path tracking and localisation errors affected by sensor contamination or failure. The focus was on the laser scanner sensor and the IMU, two major components that stop the autonomous operation. The paper looked at current technologies on the market for cleaning contaminated laser scanners and how the IMU data is currently used for impact detection. After identifying the gap, a solution for laser contamination was developed to clean a dirty laser scanner to minimise stoppages. An application was developed to utilise IMU data from the loader to detect and minimise machine damage by detecting wall impact events. The paper discussed the methods used in the development, the testing and verification of these systems. Further research into a long term replacement for the laser scanner was conducted. In the future, the IMU data use was discussed on how the autonomous algorithm could be enhanced to improve continually without degrading the environment.
|
Statistics for this ePrint Item |
Actions (login required)
 |
Archive Repository Staff Only |
