Heuristic Decision-Making for Human-aware Navigation in Domestic Environments

Kirsch, Alexandra

doi:10.29007/12l7

Cited by 2 publications

(8 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, households are more narrow and less structured than classical test environments such as museums or office spaces, and the movement should be legible 1 to people. Kirsch [15] introduces a heuristics-based local navigation method that moves a robot safely, efficiently and legibly inside single rooms, being able to circumvent obstacles without the need of a map or a planner.…”

Section: Motivationmentioning

confidence: 99%

“…Each module uses the heuristic problem solving approach from [15] (Figure 2(a) and 1), which runs in continual decision cycles. At each decision cycle a set of proposer processes compute solution alternatives.…”

Section: Heuristic Problem Solvingmentioning

confidence: 99%

“…In the case of navigation, exactly equal outcomes are very rare, therefore we always choose the (unique) best alternative. In [15], the result was executed as a control command. Here -as we have several modules -the result can either serve as input to another module or as a control command.…”

Section: Heuristic Problem Solvingmentioning

confidence: 99%

“…The experts of CONTROLLER are listed in Table 1(b), a more detailed account of the computed functions is given in [15]. The proposers comprise random sampling of commands (with only 31 commands, in the Dynamic Window Approach usual configurations use in the order of hundreds of points); the repetition of the last command; a set of "motion primitives", i.e., commands that move the robot in only one of the possible dimensions (for-/back-/sideward, turn) by a random amount; and additionally 10 commands that turn the robot on the spot.…”

Section: Modulesmentioning

confidence: 99%

“…Side-and backward movement In [15], side-/ backward movement serves as a proxy for legible, smooth navigation. In this experiment we expect little difference, possibly more such movements in classical as it may try to follow sharp curves, and slightly fewer for hps-sub (hypothesis 3).…”

Section: Datamentioning

confidence: 99%

See 4 more Smart Citations

A Modular Approach of Decision-Making in the Context of Robot Navigation in Domestic Environments

Kirsch¹

EPiC Series in Computing

Self Cite

View full text Add to dashboard Cite

Decision processes in Artificial Intelligence are often organized hierarchically. One example is robot navigation with a global path planner and a local executor. This paper examines whether a shift from optimizing the two typical modules in navigation towards a dynamic interaction of more, not necessarily hierarchically linked, modules leads to robust navigation behavior. We empirically evaluate different organizations of modules for navigation in a simulated household with a simulated PR2 robot. MotivationRobot navigation in household environments is still a challenge. Navigation in general is difficult because of the high variety of situations. In addition, households are more narrow and less structured than classical test environments such as museums or office spaces, and the movement should be legible 1 to people. Kirsch [15] introduces a heuristics-based local navigation method that moves a robot safely, efficiently and legibly inside single rooms, being able to circumvent obstacles without the need of a map or a planner.However, moving into other rooms or in very cluttered rooms requires planning. The standard approach is to calculate a global trajectory (represented by a list of intermediate points) with a global path planner and then to trace it with a local planner (also called controller) by moving the robot from one trajectory point to the next. In human-aware robot navigation, the global planner has received more research effort than the local planner [17], but the interaction of global and local planner faces some fundamental challenges: 1) When the environment changes (possibly because humans or pets are moving), a new global plan is calculated. The recalculation can be done efficiently, but the impression of goal-directed behavior is lost, the legibility of the trajectory suffers [20,16]. 2) Global planners abstract the state space and thus ignore some details. As an example, the reachability of points may be estimated by the line of sight. With a simple local planner, the planned trajectory may not be executable without * This research is funded by Deutsche Forschungsgemeinschaft and the Bavarian Academy of Sciences and Humanities.1 According to Lichtenthäler et al. [20] "Robot behavior is legible, if a human can infer the next actions, goals and intentions of the robot with high accuracy and confidence." C. Benzmüller, C.

show abstract

Section: Motivationmentioning

confidence: 99%

Section: Heuristic Problem Solvingmentioning

confidence: 99%

Section: Heuristic Problem Solvingmentioning

confidence: 99%

Section: Modulesmentioning

confidence: 99%

Section: Datamentioning

confidence: 99%

See 3 more Smart Citations

A Modular Approach of Decision-Making in the Context of Robot Navigation in Domestic Environments

Kirsch¹

EPiC Series in Computing

Self Cite

View full text Add to dashboard Cite

show abstract

Human Understanding of Robot Motion: The Role of Velocity and Orientation

Papenmeier

Uhrig

Kirsch³

2018

Int J of Soc Robotics

Self Cite

View full text Add to dashboard Cite

A general problem in human-robot interaction is how to test the quality of single robot behavior, in order to develop robust and human-acceptable skills. The most typical approach are user tests with subjective measures (questionnaires). We propose a new experimental paradigm that combines subjective measures with an objective behavioral measure, namely viewing times of images viewed as self-paced slide show. We applied this paradigm to human-aware robot navigation. With three experiments, we studied the influence of two aspects of humans' understanding of robot motion: velocity profiles and the robot's orientation. A decreasing velocity profile impaired observers' comprehension of observed motion, and robot orientations diverting from the robot's motion vector caused reduced perceived autonomy ratings. We conclude that the viewing time paradigm is a promising tool for studying humanaware robot behavior and that the design of humanaware robot navigation needs to consider both the velocity and the orientation of robots.

show abstract

Heuristic Decision-Making for Human-aware Navigation in Domestic Environments

Cited by 2 publications

References 16 publications

A Modular Approach of Decision-Making in the Context of Robot Navigation in Domestic Environments

A Modular Approach of Decision-Making in the Context of Robot Navigation in Domestic Environments

Human Understanding of Robot Motion: The Role of Velocity and Orientation

Contact Info

Product

Resources

About