Toward Reliable Off Road Autonomous Vehicles Operating in Challenging Environments

Kelly, Alonzo; Amidi, Omead; Bode, M. F.; Happold, Michael; Herman, Herman; Pilarski, Thomas; Rander, Peter; Stentz, Anthony; Vallidis, Nick; Warner, Randy

doi:10.1007/11552246_57

Cited by 70 publications

(84 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The sub-sequence of trajectories is usually computed offline [11,9]. Such methods are widely used in modern, autonomous ground robots including the two highest placing teams for DARPA Urban Challenge and Grand Challenge [28,17,27,26], LAGR [12], UPI [1], and Perceptor [14] programs. We use CONSEQOPT to maximize this function and generate trajectory sequences taking the current environment features.…”

Section: B Mobile Robot Navigationmentioning

confidence: 99%

Contextual Sequence Prediction with Application to Control Library Optimization

Dey¹,

Liu²,

Hebert³

et al. 2012

Robotics: Science and Systems VIII

View full text Add to dashboard Cite

Abstract-Sequence optimization, where the items in a list are ordered to maximize some reward has many applications such as web advertisement placement, search, and control libraries in robotics. Previous work in sequence optimization produces a static ordering that does not take any features of the item or context of the problem into account. In this work, we propose a general approach to order the items within the sequence based on the context (e.g., perceptual information, environment description, and goals). We take a simple, efficient, reduction-based approach where the choice and order of the items is established by repeatedly learning simple classifiers or regressors for each "slot" in the sequence. Our approach leverages recent work on submodular function maximization to provide a formal regret reduction from submodular sequence optimization to simple costsensitive prediction. We apply our contextual sequence prediction algorithm to optimize control libraries and demonstrate results on two robotics problems: manipulator trajectory prediction and mobile robot path planning.

show abstract

Section: B Mobile Robot Navigationmentioning

confidence: 99%

Contextual Sequence Prediction with Application to Control Library Optimization

Dey¹,

Liu²,

Hebert³

et al. 2012

Robotics: Science and Systems VIII

View full text Add to dashboard Cite

show abstract

“…To test the application of LEARCH and its modifications to learning driving styles and maneuvers, a local-global planning system was created as in [8,16], with a local planner choosing between constant curvature arcs, and a global planner operating without kinematic constraints. This planner architecture was chosen because it is simple, well understood, and in practice very effective.…”

Section: Resultsmentioning

confidence: 99%

“…Historically, the mapping of features to costs has been performed by simple manual construction of a parameterized function, and then hand tuning various parameters to create a cost function that produces desired behavior. This manual approach has been frequently used whether the cost functions in question describe locations in the world [8,16,17] or actions to be performed [4,11,19]. Unfortunately, this tedious approach typically produces subpar results, potentially leading to subpar autonomous performance.…”

Section: Related Workmentioning

confidence: 99%

“…In addition to the practical considerations mentioned above, modern planning systems are often composed of a hierarchy of individual planners, with each planner refining the results of previous plans [8,16]. In such a scenario, LEARCH must be applied to the lowest level planner, that actually makes the final decision about the next course of action.…”

Section: Learning Driving Styles and Maneuver Preferencesmentioning

confidence: 99%

See 1 more Smart Citation

Learning Autonomous Driving Styles and Maneuvers from Expert Demonstration

Silver

Bagnell

Stentz

2013

Springer Tracts in Advanced Robotics

Self Cite

View full text Add to dashboard Cite

One of the many challenges in building robust and reliable autonomous systems is the large number of parameters and settings such systems often entail. The traditional approach to this task is simply to have system experts hand tune various parameter settings, and then validate them through simulation, offline playback, and field testing. However, this approach is tedious and time consuming for the expert, and typically produces subpar performance that does not generalize. Machine learning offers a solution to this problem in the form of learning from demonstration. Rather than ask an expert to explicitly encode his own preferences, he must simply demonstrate them, allowing the system to autonomously configure itself accordingly. This work extends this approach to the task of learning driving styles and maneuver preferences for an autonomous vehicle. Head to head experiments in simulation and with a live autonomous system demonstrate that this approach produces better autonomous performance, and with less expert interaction, than traditional hand tuning.

show abstract

“…Note that log-odds values can be converted into probabilities and vice versa and we therefore store this value for each voxel instead of the occupancy probability. It is worth noting that for certain configurations of the sensor model that are symmetric, i.e., nodes being updated as hits have the same weight as the ones updated as misses, this probability update has the same effect as counting hits and misses similar to (Kelly et al, 2006).…”

Section: Probabilistic Sensor Fusionmentioning

confidence: 99%

OctoMap: an efficient probabilistic 3D mapping framework based on octrees

et al. 2013

View full text Add to dashboard Cite

Three-dimensional models provide a volumetric representation of space which is important for a variety of robotic applications including flying robots and robots that are equipped with manipulators. In this paper, we present an open-source framework to generate volumetric 3D environment models. Our mapping approach is based on octrees and uses probabilistic occupancy estimation. It explicitly represents not only occupied space, but also free and unknown areas. Furthermore, we propose an octree map compression method that keeps the 3D models compact. Our framework is available as an open-source C++ library and has already been successfully applied in several robotics projects. We present a series of experimental results carried out with real robots and on publicly available real-world datasets. The results demonstrate that our approach is able to update the representation efficiently and models the data consistently while keeping the memory requirement at a minimum.

show abstract

Toward Reliable Off Road Autonomous Vehicles Operating in Challenging Environments

Cited by 70 publications

References 9 publications

Contextual Sequence Prediction with Application to Control Library Optimization

Contextual Sequence Prediction with Application to Control Library Optimization

Learning Autonomous Driving Styles and Maneuvers from Expert Demonstration

OctoMap: an efficient probabilistic 3D mapping framework based on octrees

Contact Info

Product

Resources

About