Hidden Markov model for dynamic obstacle avoidance of mobile robot navigation

Zhu, Qiuming

doi:10.1109/70.88149

Cited by 150 publications

(67 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…There are two main techniques that fall under this category, discrete state-space techniques and clustering-based techniques. In the discrete statespace technique, the motion model is developed via Markov chains; the object state evolves from one state to another according to a learned transition probability (Zhu (2002)). In the clustering-based technique, previously-observed trajectories are grouped into different clusters, with each represented by a single trajectory prototype (Bennewitz et al (2005)).…”

Section: Related Workmentioning

confidence: 99%

Probabilistically safe motion planning to avoid dynamic obstacles with uncertain motion patterns

et al. 2013

View full text Add to dashboard Cite

This paper presents a real-time path planning algorithm that guarantees probabilistic feasibility for autonomous robots with uncertain dynamics operating amidst one or more dynamic obstacles with uncertain motion patterns. Planning safe trajectories under such conditions requires both accurate prediction and proper integration of future obstacle behavior within the planner. Given that available observation data is limited, the motion model must provide generalizable predictions that satisfy dynamic and environmental constraints, a limitation of existing approaches. This work presents a novel solution, named RR-GP, which builds a learned motion pattern model by combining the flexibility of Gaussian processes (GP) with the efficiency of RRT-Reach, a sampling-based reachability computation. Obstacle trajectory GP predictions are conditioned on dynamically feasible paths identified from the reachability analysis, yielding more accurate predictions of future behavior. RR-GP predictions are integrated with a robust path planner, using chance-constrained RRT, to identify probabilistically feasible paths. Theoretical guarantees of probabilistic feasibility are shown for linear systems under Gaussian uncertainty; approximations for nonlinear dynamics and/or non-Gaussian uncertainty are also presented. Simulations demonstrate that, with this planner, an autonomous vehicle can safely navigate a complex environment in realtime while significantly reducing the risk of collisions with dynamic obstacles.

show abstract

Section: Related Workmentioning

confidence: 99%

Probabilistically safe motion planning to avoid dynamic obstacles with uncertain motion patterns

et al. 2013

View full text Add to dashboard Cite

show abstract

“…Additionally, there are other types of models that are common in robotics, such as agent-based models (Dudek et al 1996), hidden Markov models (Zhu 1991) and state machines (Belta et al 2007). …”

Section: Model Discoverymentioning

confidence: 99%

Analyzing and improving multi-robot missions by using process mining

et al. 2017

View full text Add to dashboard Cite

Multi-robot missions can be compared to industrial processes or public services in terms of complexity, agents and interactions. Process mining is an emerging discipline that involves process modeling, analysis and improvement through the information collected by event logs. Currently, this discipline is successfully used to analyze several types of processes, but is hardly applied in the context of robotics. This work proposes a systematic protocol for the application of process mining to analyze and improve multi-robot missions. As an example, this protocol is applied to a scenario of fire surveillance and extinguishing with a fleet of UAVs. The results show the potential of process mining in the analysis of multi-robot missions and the detection of problems such as bottlenecks and inefficiencies. This work opens the way to an extensive use of these techniques in multi-robot missions, allowing the development of future systems for optimizing missions, allocating tasks to robots, detecting anomalies or supporting operator decisions.

show abstract

“…Furthermore, the robots need to be able to identify and potentially learn the intentions of people so that they can make better predictions about their future actions. In the past, various approaches have been presented to track the positions of persons (see, for example, Kluge, Köhler, and Prassler 2001;Montemerlo, Thrun, and Whittaker 2002;Schulz et al 2003) or to predict their short-term motions (Zhu 1991;Tadokoro et al 1995). These approaches assume that motion models of the persons are given.…”

Section: Introductionmentioning

confidence: 99%