Cognitive industrial floor cleaning amelioration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In industrial settings, floor cleaning is hindered by the simultaneous presence of workers and machinery, which often requires rework due to improper timing and air flow direction, leading to inefficiencies and incomplete cleaning.
Innovation Solution
A cognitive floor cleaning system that uses machine learning to analyze movement patterns of people and machines, air flow, and floor conditions to create optimized cleaning plans for robotic cleaning devices, minimizing encounters with obstacles and ensuring effective cleaning routes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If floor cleaning is performed in industrial areas with workers and machinery, then cleaning coverage is improved, but cleaning quality deteriorates due to rework from improper timing and air flow direction
Solution Approach 1:
The system performs preliminary actions by analyzing worker and machinery movement patterns before cleaning begins, predicting air flow directions, and pre-planning cleaning routes and timing to avoid rework. The machine learning model trains on historical data to anticipate future conditions, allowing the robotic cleaner to schedule and execute cleaning tasks at optimal times when areas will be less occupied and air flow conditions are favorable.
Solution Approach 2:
The system dynamically adjusts cleaning routes and timing based on real-time and predicted conditions. The machine learning model continuously learns from new data about worker patterns, machinery movements, and air flow conditions, allowing the cleaning plan to adapt and optimize itself over time. The robotic cleaner can modify its path and schedule dynamically to avoid encounters with workers and machinery while maintaining cleaning effectiveness.
2Productivity
If cleaning operations are performed in parallel with work operations, then productivity is improved, but reliability deteriorates due to worker-machinery encounters requiring rework
Solution Approach 1:
The system implements feedback mechanisms by continuously monitoring worker and machinery movement patterns, air flow conditions, and cleaning effectiveness. The machine learning model uses this feedback to refine its predictions and optimize future cleaning schedules. The system learns from each cleaning operation, adjusting its planning to improve reliability while maintaining productivity gains from parallel operations.
Solution Approach 2:
The system performs preliminary analysis of work patterns and air flow conditions before scheduling cleaning tasks. By predicting when areas will be occupied by workers or machinery and when air flow conditions will be favorable, the system can proactively schedule cleaning operations to maximize productivity while ensuring reliable completion without rework.
3Speed
If cleaning is performed against air flow direction, then cleaning speed is improved, but cleaning quality deteriorates as dust flies towards cleaned areas
Solution Approach 1:
The system performs preliminary analysis of air flow patterns in the facility before planning cleaning routes. By understanding air flow directions and speeds, the machine learning model can predict how dust will move during cleaning operations and schedule cleaning tasks at times when air flow conditions are most favorable, preventing dust from blowing onto cleaned areas while maintaining efficient cleaning speed.
4Ease of operation
If robotic cleaning devices operate autonomously, then ease of operation is improved, but device complexity increases due to machine learning models and sensor requirements
Solution Approach 1:
The system achieves universality by using a multi-functional robotic cleaning device that combines cleaning capabilities with sensing, data processing, and machine learning functions. Rather than requiring separate systems for each function, the robotic cleaner integrates multiple capabilities into a single platform, reducing overall system complexity while maintaining autonomous operation and ease of use.
Data Source
AI summary
A computer-implemented method includes receiving data regarding an area to be cleaned, the data comprising data regarding movement of people or machines through the area to be cleaned; training a machine learning cleaning model based on the data; creating a floor cleaning plan using the machine learning cleaning model, the floor cleaning plan identifying a cleaning time to minimize a likelihood of a robotic cleaning device encountering a person or a machine while cleaning; and transmitting the floor cleaning plan to a robotic cleaning device.