Applied Neural Cross-Correlation into the Curved Trajectory Detection Process for Braitenberg Vehicles
This work addresses trajectory detection for cognitive agents like robots, but it appears incremental as it builds on existing cross-correlation methods with neural mapping.
The paper tackled the problem of curved trajectory detection for Braitenberg vehicles by applying neural cross-correlation in a hard-wired circuit, achieving improved accuracy in simulated tests with a PIONEER mobile robot.
Curved Trajectory Detection (CTD) process could be considered among high-level planned capabilities for cognitive agents, has which been acquired under aegis of embedded artificial spiking neuronal circuits. In this paper, hard-wired implementation of the cross-correlation, as the most common comparison-driven scheme for both natural and artificial bionic constructions named Depth Detection Module(DDM), has been taken into account. It is manifestation of efficient handling upon epileptic seizures due to application of both excitatory and inhibitory connections within the circuit structure. Presented traditional analytic approach of the cross-correlation computation with regard to our neural mapping technique and the acquired traced precision have been turned into account for coherent accomplishments of the aforementioned design in perspective of the desired accuracy upon high-level cognitive reactions. Furthermore, the proposed circuit could be fitted into the scalable neuronal network of the CTD, properly. Simulated denouements have been captured based on the computational model of PIONEER mobile robot to verify characteristics of the module, in detail.