Robotic Cleaner Localization in Dynamic, Sensor-Blocked Spaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic cleaners face challenges in localization, particularly in dynamic environments where sensors may be blocked, leading to failure in Markov localization techniques, which assume a static environment and cannot handle multimodal distributions or uncertain initial positions.
Innovation Solution
A robotic device equipped with a camera and sensors that captures spatial data and measurements to infer its location, using a processor to generate a movement path and determine characteristics of the surroundings, such as debris accumulation and floor type, to schedule operations effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Markov localization is used with Kalman filtering, then the system can maintain a probabilistic framework for localization, but it cannot represent multimodal distributions and fails in dynamic environments where sensors are blocked
Solution Approach 1:
The patent changes the fundamental parameter of probability distribution representation from unimodal Gaussian (Kalman filtering) to multimodal distributions. This allows the system to represent multiple possible locations simultaneously, enabling recovery from sensor blockages and failures in dynamic environments while maintaining probabilistic reasoning.
Solution Approach 2:
The system transitions from assuming a static environment to handling dynamic environments where landmarks may move or be temporarily blocked. The localization algorithm dynamically adapts by maintaining multiple hypotheses about the robot's position and updating them as new sensor data becomes available, rather than relying on a single static estimate.
2Loss of information
If continuous tracking of movement is used, then the system can maintain location information, but it requires knowledge of initial location and cannot recover if tracking is lost
Solution Approach 1:
The system prepares for potential tracking loss by maintaining a map of the environment and multiple possible location hypotheses in advance. When sensor data is blocked or tracking is lost, the system can fall back on pre-computed probability distributions and environmental features to recover localization without requiring continuous unbroken tracking from an initial position.
3Ease of operation
If global localization with landmark detection is used, then the system can localize without initial position knowledge, but it fails when data contains no landmarks or in highly dynamic environments
Solution Approach 1:
The system combines multiple localization approaches into a unified framework that can handle both landmark-based global localization and continuous tracking. The multimodal probability distribution framework serves multiple functions: representing initial position uncertainty, maintaining multiple hypotheses during navigation, and recovering from sensor blockages. This makes the system universally applicable across different operational conditions.
Data Source
AI summary
Provided is a tangible, non-transitory, machine readable medium storing instructions that when executed by one or more processors of a robotic device effectuate operations including capturing, with a camera of the robotic device, spatial data of surroundings of the robotic device; generating, with the one or more processors of the robotic device, a movement path based on the spatial data of the surroundings; capturing, with at least one sensor of the robotic device, at least one measurement relative to the surroundings of the robotic device; obtaining, with the one or more processors of the robotic device, the at least one measurement; and inferring, with the one or more processors of the robotic device, a location of the robotic device based on the at least one measurement.