Subsumption Robotics
Abstract
Through the use of subsumption architectures, low-cost, simple robots can be developed to undertake the hazard of moving a submunition or scatterable mine to a disposal area for neutralization. A number of these robots acting in unison can provide an order of magnitude increase in the ability of one Explosive Ordnance Disposal (EOD) technician to clear an area. The scatterable submunitions are small in size, measuring two to five inches in maximum dimension. A two-part technique is being investigated to provide a solution to this problem. The detection of these munitions is the first step in the process. This can be achieved by modifying existing robotic vehicles with a controller, sensor, and detection processing capability. The needs for this capability are being addressed outside of this task. The second step is to provide a small, low-cost option for pick-up-carry-away (PUCA) operations for submunition clearance. Figure 1 shows prototype vehicles for the two-part technique. The long-term goal is to extend understanding of artificial intelligence principles to teams of cooperating robots, and applying the results to this real-world problem. Small autonomous vehicles can be controlled with a data-driven, task-prioritized architecture based upon subsumptive algorithms. These algorithms are used in robotics to control simple movements and tasks that would otherwise overwhelm a higher architecture controller's processing power. The primary attribute of subsumptive control is that response and motion behaviors are given to the vehicle and triggered by sensory perceptions in a data-driven manner. That is, behaviors are decided based on the application of rules to data collected by sensors. Behavior priorities are used for conflict resolution, such as encountering an object stops motion.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1998
- Accession Number
- ADA551269
Entities
People
- Anthony Healey
- Christopher K. Debolt
- Craig Freed
- George Whittinghill
- Helen Greiner
- Polly K. Pook
- Tuan N. Nguyen