Many autonomous ground vehicle (AGV) missions, such as those related to agricultural applications, search and rescue, or reconnaissance and surveillance, require the vehicle to operate in difficult outdoor terrains such as sand, mud, or snow. To ensure the safety and performance of AGVs on these terrains, a terrain-dependent driving and control system can be implemented. A key first step in implementing this system is autonomous terrain classification. It has recently been shown that the magnitude of the spatial frequency response of the terrain is an effective terrain signature. Furthermore, since the spatial frequency response is mapped by an AGV's vibration transfer function to the frequency response of the vibration measurements, the magnitude of the latter frequency responses also serve as a terrain signature. Hence, this paper focuses on terrain classification using vibration measurements. Classification is performed using a probabilistic neural network, which can be implemented online at relatively high computational speeds. The algorithm is applied experimentally to both an ATRV-Jr and an eXperimental Unmanned Vehicle (XUV) at multiple speeds. The experimental results show the efficacy of the proposed approach.

Today’s autonomous vehicles operate in an increasingly general set of circumstances. In particular, unmanned ground vehicles (UGV’s) must be able to travel on whatever terrain the mission offers, including sand, mud, or even snow. These terrains can affect the performance and controllability of the vehicle. Like a human driver who feels his vehicle’s response to the terrain and takes appropriate steps to compensate, a UGV that can autonomously perceive its terrain can also make necessary changes to its control strategy. This article focuses on the development and application of a terrain detection algorithm based on terrain induced vehicle vibration. The dominant vibration frequencies are extracted and used by a probabilistic neural network to identify the terrain. Experimental results based on iRobot’s ATRV Jr (Fig. 1) demonstrate that the algorithm is able to identify with high accuracy multi-differentiated terrains broadly classified as sand, grass, asphalt, and gravel.

No abstract

No abstract

A spatial-spectral classification technique for classification of materials within Hyperspectral images is described in this paper. The method considers the influence of neighboring pixels to apply local spatial context features to correctly label an unknown pixel. The spatial and spectral features are jointly applied to a Maximum Likelihood classifier that uses material class models defined by a Mixture of Gaussians to adaptively account for spectral variability and noise. Experimental results compare the application of spatial and spectral features with only spectral features on the classification of materials common to scenes viewed from the ground perspective.