ExcaBOT: An Autonomous Excavator for Higher Safety and Efficiency Enabled by Digital Technologies
Virtual Vehicle Research Center
Every year in the construction industry all over the world, people are killed or injured as a result of being struck by moving construction machines. Vehicle reversing operations cause a third of all fatal transport accidents in the construction industry, producing an average of five deaths and 20 major injuries per year. To provide safety while enhancing the operational efficiency, digital technologies should be used. Autonomy and remote operation of construction machinery in construction sector has a great potential transferring manufacturing industry digitalization gains (Industry 4.0) to the construction industry. Autonomous and tele-operated construction equipment can be operated with little or no human intervention for performing specified tasks in high safety and efficiency. SmartUniversal proposal is to enable safe and accurate control of excavators (tracks and arms) operating in autonomous modes, avoiding collisions with obstacles, other machines, or people in the construction site thanks to sensor fusion and machine vision technologies. This experiment will accelerate the uptake of digital technologies in the construction area by introducing and demonstrating autonomous operation of excavators in real environment.
- Challenge 1: Excavators present hazards, such as restricted visibility and lack of driver protection from the effects of overturning, noise, and vibration. These are even more severe for older excavator models. Materials handling vehicles such as excavators are at risk of overturning when overloaded, working on sloping, uneven or unstable ground or driving with raised loads.
- Challenge 2: Stability under all foreseeable operating conditions with lack of human error
- Challenge 3: Adequate visibility for the operator all around the vehicle (operator only checking from outside)
- Challenge 4: Protection for the operator from hazards, such as working at height, falling objects, and overturning
- Challenge 5: Protection of operator from the weather, noise, vibration, noxious fumes, and dusts.
In our project, we have demonstrated autonomously working excavators for specific excavation tasks for optimized excavation work in the most efficient way (high excavation in a short time) without jeopardizing operational safety. To prove the concept, we demonstrated a real-world use case that not only involves detecting and recognizing the construction workers on the test site using perception methods and associated sensors such as camera and LIDAR, but also enables more efficient excavation operations. The security system generates a hard warning signal when the construction workers enter the predetermined unsafe area in dangerous proximity.
Starting with reviewing the system design on a 1:14 scale excavator, the desire to think bigger arose. To demonstrate the autonomous operation of an excavator, we had:
- Demonstrated an existing control strategy for autonomous excavator operation from the laboratory (TRL4) to the real excavator in the real construction environment (TRL7).
- The control strategy of autonomous excavators was successfully verified on a real excavator using the independend co-simulation framework ICOS from DIH Virtual Vehicle, due to previous simulation tool was insufficient.
- The autonomous excavation could be demonstrated on a test site on a sensor-based and automated standard heavy-duty excavator from HİDROMEK.
We used sensor fusion for LIDAR, stereo/cameras for autonomous operation of the excavator to move forward and backward, dig earth and avoid all collisions with other machines and people around (any object bigger and higher than is a traffic cone, within 5 meters range)
The next challenge will be to enhance the system towards full autonomy with knowledge and experience gained from ExcaBOT.
LinkedIn post with a video of the successful demonstration: