Assessing terrain ahead of a robot when repeating previously driven safe paths can be accomplished by looking for geometric changes (e.g., due to the appearance of humans or other obstacles). Previous work has shown that the incorporation of data-driven learning and place-dependence are useful aspects of making terrain classification viable in challenging terrain. This paper presents a learning, place-dependent (LPD) terrain classifier that uses a probabilistic model of the terrain to improve detection of small obstacles in uncluttered terrain while avoiding false positives in more challenging environments. Specifically, a Gaussian mixture model is used to account for multi-height terrain cells that arise in heavily vegetated areas (where both a ground plane and overhanging vegetation can occupy the same cell). A variational Bayesian technique is used to automatically determine the number of components required for each cell using a Dirichlet prior on mixing proportions and a Normal-Inverse-Wishart prior on the means and covariances of the components. The probabilistic nature of the model allows for the detection of much smaller obstacles in regions that exhibit low variance in the terrain surface, whilst still avoiding false positives in regions where the terrain is highly cluttered (e.g., vegetation). The algorithm is tested on almost 10 km of autonomous traverse and is shown to be able to classify a wider range of obstacles than two baseline change-detection algorithms based on absolute geometric differences.

Autonomous path-following systems based on the Teach and Repeat paradigm allow robots to traverse extensive networks of manually driven paths using on-board sensors. These methods are well suited for applications that involve repeated traversals of constrained paths such as factory floors, orchards, and mines. In order for path-following systems to be viable for these applications they must be able to navigate large distances over long time periods, a challenging task for vision-based systems that are susceptible to appearance change. This paper details Visual Teach and Repeat 2.0, a vision-based path-following system capable of safe, long-term navigation over large-scale networks of connected paths in unstructured, outdoor environments. These tasks are achieved through the use of a suite of novel, multi-experience, vision-based navigation algorithms. We have validated our system experimentally through an eleven-day field test in an untended gravel pit in Sudbury, Canada, where we incrementally built and autonomously traversed a 5 Km network of paths. Over the span of the field test, the robot logged over 140 Km of autonomous driving with an autonomy rate of 99.6%, despite experiencing significant appearance change due to lighting and weather, including driving at night using headlights.

This paper presents a learned, place-dependent terrain-assessment classifier that improves over time. Whereas typical methods aim to assess all of the terrain in a given environment, we exploit the fact that many robotic navigation tasks are well-suited to visual-teach-and-repeat navigation where robot motion is restricted to previously driven paths. In such scenarios, we argue that general terrain assessment is not required, and we can instead solve the much easier problem of detecting changes along the path; we sacrifice the ability to generalize off the path in favour of improved performance on the path. Terrain along a pretaught path is compared to terrain seen at the same location during previous, human-supervised traverses, and any significant differences cause that location to be labelled as unsafe. By storing all of our previous experiences, we are able to continuously improve our estimates as we revisit the same locations multiple times. We tested our method on two datasets collected at the University of Toronto and show that we improve over existing place-independent (both learned and not) methods, enabling nearly full autonomy in challenging, varied terrain using only a stereo camera.

This paper presents an approach to learning robot terrain assessment from human demonstration. An operator drives a robot for a short period of time, supervising the gathering of traversable and untraversable terrain data. After this initial training period, the robot can then predict the traversability of new terrain based on its experiences. We improve on current methods in two ways: first, we maintain a richer (higher-dimensional) representation of the terrain that is better able to distinguish between different training examples. Second, we use a Gaussian-process classifier for terrain assessment due to its superior introspective abilities (leading to better uncertainty estimates) when compared to other classifier methods in the literature. Our method is tested on real data and shown to outperform current methods both in classification accuracy and uncertainty estimation.

We have been developing a navigational framework called Network of Reusable Paths that enables a rover to revisit any previously driven-to location quickly and cheaply. This place-revisiting capability enables the parallel analysis of scientific data from several sites of interest in planetary exploration missions, which allows for a methodical downselection of the locations to determine the best candidate site for costly scientific operations (e.g., sampling). The results from an analogue mission at the Canadian Space Agency’s Mars Emulation Terrain in Montréal, Canada are presented in which a 220 metre network is built by teleoperating the rover under a five second communication delay. The rover is operated from a remote backroom consisting of science subteams (each assigned to a site) and a rover operations team.

This paper presents a proof-of-concept, rover-based system to locate the source(s) of methane gas on Mars. A distributed open-path spectrometer is mounted on a rover and pointed at several retroreflective signs to measure the line-of-sight methane concentration between the rover and sign. By moving the rover around and accumulating such measurements, the location of a methane source can be determined. The system was integrated and tested over a 2 kilometre traverse in an analog mission at the Canadian Space Agency on the Mars Emulation Terrain. Engineering results suggest that total driving distance is comparable to other similar methods, but by using a network of reusable paths, the total distance driven in new terrain can be reduced by a factor of five. Autonomous pointing of signs was successful 80% of the time. Future work includes combining the engineering and science data from the analog mission to better understand how to locate methane sources.

Planetary rovers to date have been operated mainly in a serial mode; they are driven from one place to the next, away from the lander, and seldom return to previously visited places. We have been developing a visual navigation technique, called network of reusable paths (NRP), that can be thought of as a low-computational-cost version of simultaneous localization and mapping coupled to a path-tracking controller. The result is that a rover can be returned accurately to any place it has previously visited using only visual feedback; this enables science to be gathered from multiple sites in parallel. We will describe how NRP works, and present field test results of two mission concepts where we have made use of this technology: a lunar-sample-return scenario and a Mars-methane-hunting scenario.