VALIDATION OF A SIMULATED ENVIRONMENT DEVELOPED FOR VALIDATING CFS AUTONOMY
Ruel Ellis1, Gerardo Fernandez-Lopez2 and Gerard Pounder3*
1Faculty of Engineering, The University of the West Indies, Trinidad
2Faculty of Engineering, Simón Bolivar University, Venezuela
3Faculty of Engineering, The University of the West Indies, Trinidad
1Email: ruel.ellis@sta.uwi.edu
2Email: gerardo.fernandez.lopez@gmail.com
3Email: pounderji@yahoo.co.uk *(Corresponding author)
Abstract:
This paper analyses results from experiments performed using a previously-described, simulated environment that was developed for validation of Cognitive Function Synthesis, or CFS, Autonomy. Navigation performance of the Pioneer robot platform used the following metrics: Average Cycle Time per simulation run; Average Wall Contact per cycle; and Average Shock Treatment Activation per simulation run. Two ultrasound, or US, configurations were used while the robot navigated in either the ‘preconfigured-reflexes only’ mode or the ‘Braitenberg Obstacle Avoidance’ mode. Results from the “16 sensor” US configuration was generally found to be significantly different from that of the “8 sensor” configuration, independently of obstacle avoidance considerations. Robot performance, when subject to the Braitenberg Obstacle Avoidance algorithm, was also found to be significantly different from ‘preconfigured-reflexes only’ performance, regardless of US configuration. The difference in Shock Treatment and the Average Wall Contact, observed between the “16 sensor” US setting and the “8 sensor” configuration for the ‘Braitenberg Obstacle Avoidance’, are likely to be due to the coefficient values adopted for the rear US sensors together with robot position at experiment start. The use of this environment to enable statistical analysis of results, to determine significant difference in obstacle avoidance performance, validates its usefulness as a tool for CFS Autonomy validation.
Keywords: Cognitive Function Synthesis, Artificial General Intelligence, Associative Memories, Autonomous Navigation, Biomimetic Navigation.
https://doi.org/10.47412/EBSY1245
