A GPU-accelerated cortical neural network model for visually guided robot navigation

Abstract

Humans and other terrestrial animals use vision to traverse novel cluttered environments with apparent ease. On one hand, although much is known about the behavioral dynamics of steering in humans, it remains unclear how relevant perceptual variables might be represented in the brain. On the other hand, although a wealth of data exists about the neural circuitry that is concerned with the perception of self-motion variables such as the current direction of travel, little research has been devoted to investigating how this neural circuitry may relate to active steering control. Here we present a cortical neural network model for visually guided navigation that has been embodied on a physical robot exploring a real-world environment. The model includes a rate based motion energy model for area V1, and a spiking neural network model for cortical area MT. The model generates a cortical representation of optic flow, determines the position of objects based on motion discontinuities, and combines these signals with the representation of a goal location to produce motor commands that successfully steer the robot around obstacles toward the goal. The model produces robot trajectories that closely match human behavioral data. This study demonstrates how neural signals in a model of cortical area MT might provide sufficient motion information to steer a physical robot on human-like paths around obstacles in a real-world environment, and exemplifies the importance of embodiment, as behavior is deeply coupled not only with the underlying model of brain function, but also with the anatomical constraints of the physical body it controls.

A Directionally Selective Small Target Motion Detecting Visual Neural Network in Cluttered Backgrounds

“…In the first experiment, (n 4 , τ 4 ) is set as (1,5), (2,10), (3,15), (4,20), (5,25), (6,30). In the second experiment, (n 5 , τ 5 ) is set as (3,15), (4,20), (5,25), (6,30), (7,35), (8,40). In the third experiment, (σ 4 , σ 5 ) is set as (1.0, 2.0), (1.5, 3.0), (2.3, 4.6), (2.8, 5.6), (3.7, 7.4).…”

“…Tuning properties of the proposed neural network under different parameter (n 5 , τ 5 ). In this experiment, (n 5 , τ 5 ) is set as(3,15),(4,20),(5,25),(6,30),(7,35),(8,40) while the other parameters are fixed. In each subplot, the horizontal axis represents one of the target parameters (Weber Contrast, velocity, width, and height) while the vertical axis denotes normalized model outputs.…”

Exploration and Mining Learning Robot of Autonomous Marine Resources Based on Adaptive Neural Network Controller

“…Taking the neural network as the starting point, the nonlinear system identification method based on neural network and the network characteristics commonly used for identification are discussed. Study showed the neural network is optimized from two aspects of network topology and learning algorithm [4]. The neural network identification model of multi-layer forward neural network and dynamic regression neural network based on improved Elman algorithm is studied.…”

Age estimation of Calliphora (Diptera: Calliphoridae) larvae using cuticular hydrocarbon analysis and Artificial Neural Networks

“…DNA- 28 based analyses have been applied to the field of forensic entomology for over a decade [6]. They 29 have been successfully used to identify and age forensically important species [7][8][9][10][11]. 30 One technique that may have the potential to give the same accuracy as DNA-based techniques for 31 an ageing tool is Cuticular Hydrocarbon (CHC) analysis [12].…”

“…and have the ability to learn the relationship between input and output data, making them well-55 suited to fast processing of large real-world datasets (e.g. speech recognition [28], robotics [29], 56 structural engineering [30] as well as the forensic sciences [31].…”