MIT News breed · Wetenschap
MIT Engineers Develop Intuitive Controller for Excavators
MIT engineers have designed a novel training interface for excavator operators that replaces traditional joysticks with a miniature mechanical arm and bucket, aiming to significantly shorten the learning curve for operating heavy machinery.
Operating an excavator, much like a claw machine, requires complex mental mapping to coordinate controls like joysticks with the machine's movements. This process can take years to master. The new interface, called the "World-Space Interface" (WSI), mimics the excavator's arm and bucket, allowing trainees to control a virtual excavator on an immersive display by physically manipulating the device.
Hermano Krebs, a principal research scientist at MIT, stated that the WSI offers a more intuitive way to command the machine, potentially eliminating the need for operators to develop complex mental maps. He envisions the interface not only as a training tool but also as a direct control mechanism, either within the excavator cab as an exoskeleton-like appendage or for remote teleoperation in hazardous environments.
The collaboration with Sumitomo Heavy Industries stemmed from a need to train new operators as Japan's heavy machinery workforce ages. Traditional training involves learning to correlate joystick movements with the excavator's actions, a process that can take months to years. The WSI aims to bypass this by allowing operators to directly mimic the machine's movements.
In comparative training experiments, novices using the WSI performed as well as experts from the start, unlike those using traditional joystick simulators who showed a larger gap between novice and expert performance. The team is now working to incorporate haptic feedback into the WSI to provide a sense of touch, further enhancing its intuitiveness.
The WSI presents a more natural alternative to existing virtual simulators developed by companies like Caterpillar and Hyundai, which are largely based on conventional joystick controls. The MIT team believes their interface could enable even first-time operators to work effectively from day one.
AI-samenvatting op basis van de bron.
MIT News breed