Surface Detection Classifier for Micromobility Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Micromobility vehicles often operate on unintended surfaces, interfering with pedestrians and other travelers, which reduces their appeal in urban areas.
Innovation Solution
A surface detection classifier is trained to identify restricted surfaces using sensor data and tagged image data, with the system notifying users, limiting speed, or disabling the vehicle if it detects operation on such surfaces for a threshold period.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If surface detection and control measures are implemented to prevent operation on restricted surfaces, then interference with pedestrians and other travelers is reduced, but device complexity increases
Solution Approach 1:
The patent replaces complex mechanical surface detection systems with sensor-based detection (accelerometers, gyroscopes, magnetometers) and machine learning classifiers. The sensor suite detects vehicle movement characteristics and the classifier determines surface type through pattern recognition, eliminating the need for complex mechanical sensors while achieving accurate surface detection to prevent pedestrian interference.
Solution Approach 2:
The patent introduces a machine learning classifier as an intermediary between raw sensor data and surface type determination. This classifier processes sensor inputs and outputs surface type classifications, serving as a mediator that simplifies the overall system architecture while enabling accurate detection of restricted surfaces to prevent interference with pedestrians and other travelers.
2Measurement precision
If sensor data collection and machine learning classification are used to identify restricted surfaces, then detection accuracy is improved, but loss of time in data processing increases
Solution Approach 1:
The patent performs preliminary action by pre-training the machine learning classifier offline with extensive sensor data from various surfaces. During actual operation, the pre-trained classifier quickly processes sensor inputs without requiring real-time training, thus achieving high detection accuracy while minimizing processing time during vehicle operation.
Solution Approach 2:
The patent collects excessive sensor data (acceleration, gyroscope, magnetometer readings at multiple frequencies) beyond the minimum required for basic detection. This excessive data collection provides redundant information that improves classification accuracy through pattern recognition, while the system processes only the essential features extracted from this data, balancing accuracy with processing efficiency.
Data Source
AI summary
Various implementations include approaches for training a surface detection classifier and detecting characteristics of a surface, along with related micromobility vehicles. Certain implementations include a method including: comparing: i) detected movement of a micromobility (MM) vehicle or a device located with a user at the MM vehicle while operating the MM vehicle, with ii) a surface detection classifier for the MM vehicle; and in response to detecting that the MM vehicle is traveling on a restricted surface type for a threshold period, performing at least one of: a) notifying an operator of the MM vehicle about the travel on the restricted surface type, b) outputting a warning at an interface connected with the MM vehicle or the device, c) limiting a speed of the MM vehicle, or d) disabling operation of the MM vehicle.