Crab-like legged robot for accessing and classifying munitions in surf zones
Dr. Kathryn Daltorio | Case Western Reserve University
Unexploded ordinance (UXO) from military training are potential environmental hazards that can be embedded in mud, sand and rocky terrain. Especially in shallow water (less than 5m), these munitions are difficult to detect and access with current autonomous vehicles.
The long-term goal of this research is to develop small amphibious vehicles that can make surf zones safer by retrieving, disabling, and marking the locations of munitions within fields of other debris. These robots could be deployed from ships or docks, replacing human divers. This project will examine crab-like robot legs for traversing surf zones to classify objects partially-buried in the terrain with leg-embedded force sensors. Inspired by crabs, the novel concept proposed is to use compliance to grasp the ground during destabilizing waves and release to “punt” when hydrodynamic currents are favorable. The ability to selectively anchor to the ground will be useful for many operations underwater (e.g, grasping small munitions or stabilizing close to the ground for sensor scans).
This project will (1) demonstrate a planar legged walking platform for grasping terrain, (2) develop sensor-embedded legs for tactile classification of terrain and objects, and (3) evaluate the robot in lab tests and field tests at Lake Erie beaches.
Animals like crabs can traverse sandy, rocky, and wave-swept wet and dry terrain that current robots cannot cross. The hypothesis is that by selectively creating inward tangential forces between opposing legs, crabs grasp the ground for traction greater than their weight when needed to be able to anchor during waves and then release for efficient wave-riding and swimming.
This project will develop proof-of-concept walking gaits that incorporate grasping for surf zones. This requires only a simple planar robot platform. Klann mechanism legs allow each leg to be controlled by a single rotating actuator (four servomotors total). Compliance will allow passive grasping with an open-loop gait, which is expected to be effective over uniform granular media. Sensors embedded in the legs will enable a responsive gait to take advantage of hydrodynamics and search for footholds among rocks. These sensors will also provide tactile data on terrain geometry (roughness, convexity, slope) and forces (stiffness, stickiness, friction) that will help classify partially buried munitions (simulated by metal cylinders for planar tests). The project will test the effectiveness, sufficiency, and future requirements of these behaviors in a small wave tank in a lab and at Great Lakes beaches.
The anticipated outcome is a robot for using waves, rocks, and sand for quiet, efficient locomotion in close range surveys between sea beds and dry land. This project has the potential to inform future efforts to apply crab-like locomotion for environmental tools, especially munitions response.
Fully developed crab-robots will navigate autonomously. They will include specialized sensors and appendages to help the Department of Defense evaluate hazards in critical coastal zones, facilitate rapid retrieval of overboard objects, and perform amphibious repairs or other missions to reduce environmental impact of military operations. In robotics, a wave-riding crab-like gait would be a novel scientific achievement. In machine learning, predicting periodic wave currents would be a new application for tactile sensing. For biology, this robot would be a physical model of arthropod walking, which could be used to test biologically-inspired controllers that have been developed based on animal experiments. Environmentally, such a robot could stealthily observe marine life.