A novel floor segmentation algorithm for mobile robot navigation

Bhowmick, Soumabha; Pant, Abhishek; Mukherjee, Jayanta; Deb, Alok Kanti

doi:10.1109/ncvpripg.2015.7490000

Cited by 7 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Procedure data collection. BFS is a common path planning algorithm for navigation [77,78,79,80], which we use to generate expert trajectories for training an imitation learning agent. To compute the optimal action at each time step, BFS keeps track of a visited 2D array (colored cells in Figure 3) that marks whether each position (1) has been visited by the search, (2) if so, which action visited it, and (3) has a position been backtracked.…”

Section: Proof Of Concept: Synthetic Maze Navigationmentioning

confidence: 99%

Chain of Thought Imitation with Procedure Cloning

Yang¹,

Schuurmans²,

Abbeel³

et al. 2022

Preprint

View full text Add to dashboard Cite

Imitation learning aims to extract high-performance policies from logged demonstrations of expert behavior. It is common to frame imitation learning as a supervised learning problem in which one fits a function approximator to the input-output mapping exhibited by the logged demonstrations (input observations to output actions). While the framing of imitation learning as a supervised input-output learning problem allows for applicability in a wide variety of settings, it is also an overly simplistic view of the problem in situations where the expert demonstrations provide much richer insight into expert behavior. For example, applications such as path navigation, robot manipulation, and strategy games acquire expert demonstrations via planning, search, or some other multi-step algorithm, revealing not just the output action to be imitated but also the procedure for how to determine this action. While these intermediate computations may use tools not available to the agent during inference (e.g., environment simulators), they are nevertheless informative as a way to explain an expert's mapping of state to actions. To properly leverage expert procedure information without relying on the privileged tools the expert may have used to perform the procedure, we propose procedure cloning, which applies supervised sequence prediction to imitate the series of expert computations. This way, procedure cloning learns not only what to do (i.e., the output action), but how and why to do it (i.e., the procedure). Through empirical analysis on navigation, simulated robotic manipulation, and game-playing environments, we show that imitating the intermediate computations of an expert's behavior enables procedure cloning to learn policies exhibiting significant generalization to unseen environment configurations, including those configurations for which running the expert's procedure directly is infeasible. 1

show abstract

Section: Proof Of Concept: Synthetic Maze Navigationmentioning

confidence: 99%

Chain of Thought Imitation with Procedure Cloning

Yang¹,

Schuurmans²,

Abbeel³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…However, since pixel-based evaluation and edge filters were emphasized in the study, objects that were very similar to the floor, such as the wall could be indistinguishable. Bhowmick et al [18] relied on floor detection using a conventional breadth-first search-based region-growing technique to detect obstacles. However, they stated that the process was slow because of the many conditions in the study.…”

Section: Literature Reviewmentioning

confidence: 99%

Real-Time Obstacle Avoidance Based on Floor Detection for Mobile Robots

Hiçdurmaz

Tuncer

2020

Sakarya University Journal of Science

View full text Add to dashboard Cite

Obstacle detection and avoidance are two main problems that demand solutions in the autonomous movement of mobile robots. To this end, the robots have been equipped with sensors and cameras. This study proposes a new method that allows mobile robots to move freely without any collision in an uncertain (i.e., both static and dynamic) workspace by processing images taken using a real-time webcam. In the study, a robot was allowed to move depending on the visibility and suitability of the floor in the images. These steps were repeated for each new image and, furthermore, the images were segmented based on an adaptive threshold obtained by calculating the statistical parameters. This segmentation was aimed to separate the floor from other areas in the study. Experimental results demonstrate that the proposed method is extremely successful to separate the floor from other regions and has a low cost and flexible method for obstacle avoidance.

show abstract

An interactive robot design to find missing people and inform their location by real-time face recognition system on moving images

Avuçlu

Başçiftçi

2021

J Ambient Intell Human Comput

View full text Add to dashboard Cite

A novel floor segmentation algorithm for mobile robot navigation

Cited by 7 publications

References 17 publications

Chain of Thought Imitation with Procedure Cloning

Chain of Thought Imitation with Procedure Cloning

Real-Time Obstacle Avoidance Based on Floor Detection for Mobile Robots

An interactive robot design to find missing people and inform their location by real-time face recognition system on moving images

Contact Info

Product

Resources

About