This paper describes the motivation, design and implementation of a Perception Validation System (PVS), a system for measuring the outdoor perception performance of an autonomous vehicle. The PVS relies on using large amounts of real world data and ground truth information to quantify performance aspects such as the rate of false positive or false negative detections of an obstacle detection system. Our system relies on a relational database infrastructure to achieve a high degree of flexibility in the type of analyses it can support. We discuss the main steps required for going from raw data to numerical estimates describing the performance of the perception system, including the generation of ground truth information and the safe speed metric we found to be most useful for comparing the perception system's outputs to the ground truth data. We present results illustrating some of the analyses that can be completed using the Perception Validation System.

Autonomous navigation in outdoor, off-road environments requires solving complex classification problems. Obstacle detection, road following and terrain classification are examples of tasks which have been successfully approached using supervised machine learning techniques for classification. Large amounts of training data are usually necessary in order to achieve satisfactory generalization. In such cases, manually labeling data becomes an expensive and tedious process. This work describes a method for reducing the amount of data that needs to be presented to a human trainer. The algorithm relies on kernel density estimation in order to identify "interesting" scenes in a dataset. Our method does not require any interaction with a human expert for selecting the images, and only minimal amounts of tuning are necessary. We demonstrate its effectiveness in several experiments using data collected with two different vehicles. We first show that our method automatically selects those scenes from a large dataset that a person would consider "important" for classification tasks. Secondly, we show that by labeling only few of the images selected by our method, we obtain classification performance that is comparable to the one reached after labeling hundreds of images from the same dataset.

This work describes an approach for using several levels of data fusion in the domain of autonomous off-road navigation. We are focusing on outdoor obstacle detection, and we present techniques that leverage on data fusion and machine learning for increasing the reliability of obstacle detection systems. We are combining color and infrared (IR) imagery with range information from a laser range finder. We show that in addition to fusing data at the pixel level, performing high level classifier fusion is beneficial in our domain. Our general approach is to use machine learning techniques for automatically deriving effective models of the classes of interest (obstacle and non-obstacle for example). We train classifiers on different subsets of the features we extract from our sensor suite and show how different classifier fusion schemes can be applied for obtaining a multiple classifier system that is more robust than any of the classifiers presented as input. We present experimental results we obtained on data collected with both the experimental unmanned vehicle (XUV) and a CMU developed robotic tractor.

Describes the design and implementation of an algorithm for improving the performance of stereo vision in environments presenting repetitive patterns or regions with relatively weak texture. The proposed algorithm makes use of the common assumption that the disparities corresponding to continuous surfaces in the world vary smoothly; we use this assumption to alleviate the correspondence problem for pixels that cannot be reliably matched by the stereo algorithm. Our approach can be described as a reliability based filtering of the disparity image followed by a recursive propagation step. It can be applied to the output of almost any "standard" stereo algorithm with minimal modifications, and is computationally efficient.

We propose a new system for vision-based mobile robot navigation in an unmodeled environment. Simple, unobtrusive artificial landmarks are used as navigation and localization aids. The landmark patterns are designed so that they can be reliably detected in real-time in images taken with the robot's camera over a wide range of viewing configurations. The code for the recognition algorithm is available in the web site.

No abstract available.