Search

Mobile robots are often tasked with repeatedly navigating through an environment whose traversability changes over time. These changes may exhibit some hidden structure, which can be learned. Many studies consider reactive algorithms for online planning, however, these algorithms do not take advantage of the past executions of the navigation task for future tasks. In this article, we formalize the problem of minimizing the total expected cost to perform multiple start-to-goal navigation tasks on a roadmap by introducing the learned reactive planning problem. We propose a method that captures information from past executions to learn a motion policy to handle obstacles that the robot has seen before. We propose the LAMP framework, which integrates the generated motion policy with an existing navigation stack. Finally, an extensive set of experiments in simulated and real-world environments show that the proposed method outperforms the state-of-the-art algorithms by 10% to 40% in terms of expected time to travel from start to goal. We also evaluate the robustness of the proposed method in the presence of localization and mapping errors on a real robot. [ABSTRACT FROM AUTHOR]

Objective: We determined the rate of, and predictive factors for, subsequent impaired driving activity (IDA) by injured drivers treated in a Canadian tertiary care emergency department (ED) following a motor vehicle crash (MVC). Methods: We retrospectively identified all drivers injured in a MVC who presented to our tertiary care, urban ED (1999-2003) and had their blood alcohol content (BAC) measured. Injured drivers were categorized by BAC: group 1, BAC = 0; group 2, 0 < BAC ≤ 17.3 mM (80 mg/dL, legal limit); and group 3, BAC > 17.3 mM. IDA was defined as any of the following: a conviction for impaired driving; a 24-h or 90-day license suspension for impaired driving; involvement in alcohol-related MVC. Time to IDA following the index event between groups was compared with Kaplan-Meier survival analyses. Effects of covariates on time to IDA were analyzed using Cox proportional hazards models. Results: During the study period, 1489 injured drivers met study criteria: 1171 in group 1, 51 in group 2, and 267 in group 3. During an average follow-up of 52.4 months, 82 (30.7%) group 3 drivers engaged in subsequent IDA, compared with 80 (6.8%) group 1 drivers (p < 0.0001). Youth, male gender, history of previous IDA, and the number of previous IDA events were all associated with a significant increase in subsequent IDA. A history of IDA was the strongest predictor of future IDA in group 1 (440% increase risk) and in group 3 (80% increased risk). The magnitude of BAC elevation above the legal limit was not predictive of future IDA. Conclusions: A high portion of injured impaired drivers who present to hospital engage in repeat IDA following discharge. Besides impairment at time of hospital visit, the best predictor of future IDA is a history of IDA prior to the index event. [ABSTRACT FROM AUTHOR]