An Autonomous Robot-Aided Auditing Scheme for Floor Cleaning

Abstract

Cleaning is an important factor in most aspects of our day-to-day life. This research work brings a solution to the fundamental question of “How clean is clean” by introducing a novel framework for auditing the cleanliness of built infrastructure using mobile robots. The proposed system presents a strategy for assessing the quality of cleaning in a given area and a novel exploration strategy that facilitates the auditing in a given location by a mobile robot. An audit sensor that works by the “touch and inspect” analogy that assigns an audit score corresponds to its area of inspection has been developed. A vision-based dirt-probability-driven exploration is proposed to empower a mobile robot with an audit sensor on-board to perform auditing tasks effectively. The quality of cleaning is quantified using a dirt density map representing location-wise audit scores, dirt distribution pattern obtained by kernel density estimation, and cleaning benchmark score representing the extent of cleanliness. The framework is realized in an in-house developed audit robot to perform the cleaning audit in indoor and semi-outdoor environments. The proposed method is validated by experiment trials to estimate the cleanliness in five different locations using the developed audit sensor and dirt-probability-driven exploration.

A Reinforcement Learning Based Dirt-Exploration for Cleaning-Auditing Robot

“…Then, based on the policy it learned, the robot chooses the locations it has to go to in the rollout phase. The proposed method enhances the space-auditing procedure in the cleaning-auditing framework [22], since its performance depends on training data which can be simulated easily. The overall objective of this research work is subdivided into four components.…”

“…To start with the proposed strategy for auditing, the background of the audit robot mentioned in Pathmakumar et al [22] (also known as BELUGA robot) for completeness and clarity. The Figure 1 shows the overview of components and procedures involved in cleaning auditing.…”

“…Leveraging upon the advances in robotic technology, developing an autonomous cleaning-auditing robot is an effective solution to address the problem. The primary footstep to address this problem was reported in our previous research work mentioned in Pathmakumar et al [22]. The above mentioned research work introduces a framework for cleaning-auditing using an autonomous mobile robot with a sample audit sensor as a critical payload.…”

sTetro-Deep Learning Powered Staircase Cleaning and Maintenance Reconfigurable Robot

“…Instead of traditional computer vision based techniques, learning based environmental perception (e.g., Deep Learning) is a new paradigm. The deep learning neural network layers can learn features to distinguish and determine objects at high accuracy and can be used in various vision tasks such as outdoor environmental perception for autonomous vehicles [ 18 , 19 , 20 ], cleaning and maintenance robot vision pipeline [ 2 , 21 , 22 , 23 , 24 ], mobile robot place recognition and mapping tasks [ 25 , 26 , 27 , 28 ]. While these state-of-the-art deep learning algorithms have achieved impressive results for environment perception design, there are some shortcomings such as false detection, which can be due to objects with similar features.…”

Drain Structural Defect Detection and Mapping Using AI-Enabled Reconfigurable Robot Raptor and IoRT Framework

“…Reconfigurable robots are more flexible than fixed-size robots. In recent years, reconfigurable morphology-based robots have been widely used for various automation applications including cleaning indoor and outdoor environments [25,26], inspecting built environments [23,27], ship hull inspection, aircraft structural inspection [28], etc. From the literature, it is realized that designing a reconfigurable morphology-based robot is an optimal solution to overcoming the shortcomings mentioned earlier.…”