Autonomous Lawn Mower Boundary Detection via Camera and Grass Sensor
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous lawn mowers face challenges in reliably distinguishing between usable and non-usable areas, leading to inefficient and error-prone mowing operations due to limitations in detection technologies, such as point-effect grass sensors that can misclassify surfaces.
Innovation Solution
The integration of an electronic image recording unit with a camera and a surface sensor that uses probability distribution to classify pixels, allowing for accurate area discrimination and adjustment of sensor signals to improve detection quality, enabling precise movement control and operation within defined boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a point-effect grass sensor is used for area discrimination, then the device complexity is reduced, but the measurement precision deteriorates leading to incorrect surface classification
Solution Approach 1:
The patent divides the detection task into two segments: a point-effect grass sensor detects specific grass characteristics at a focused point, while an electronic image recording unit (camera) captures broader contextual information. This segmentation allows each sensor type to contribute its strengths - the grass sensor provides precise spectral analysis while the camera provides spatial context, together resolving the contradiction between simple detection and accurate classification
Solution Approach 2:
The evaluation unit acts as an intermediary that processes and combines signals from both the grass sensor and the electronic image recording unit. It integrates the point-effect spectral data with the broader visual context, mediating between the simple detection capability and the need for precise surface classification by synthesizing multiple information sources
2Reliability
If a boundary wire system is used to mark the processing area, then the reliability of area delimitation is improved, but the device complexity and ease of operation worsen due to installation complexity and stochastic control requirements
Solution Approach 1:
The patent extracts the boundary detection function from the physical boundary wire system and transfers it to the sensor system. Instead of requiring a wire to be laid around the lawn boundary, the grass sensor and camera detect the boundary naturally through spectral and visual differences between lawn and non-lawn areas. This removes the complex wire installation while maintaining reliable area delimitation
Solution Approach 2:
The system uses the natural optical and spectral properties of the lawn surface itself to define boundaries, rather than requiring external markers. The grass sensor detects the spectral signature of grass to identify usable areas, and the camera provides contextual verification, allowing the environment to serve its own boundary-marking function without additional infrastructure
3Ease of operation
If a selective grass sensor is used for detection, then the ease of operation is improved, but the measurement precision deteriorates due to point-effect detection leading to incorrect conclusions about surface character
Solution Approach 1:
The patent merges the outputs of the grass sensor and the electronic image recording unit in the evaluation unit. The grass sensor provides spectral information about grass presence, while the camera provides visual context about the broader surface area. By combining these complementary data sources, the system maintains simple operation while achieving accurate surface character detection that neither sensor could achieve alone
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This solution significantly enhances the accuracy and reliability of area discrimination, ensuring precise mowing within boundaries and preventing unnecessary operation on non-usable areas, thereby improving processing quality and adherence to demarcation.
Implementation Method 1
the evaluation of the spectral properties of the daylight or the sun of the light reflected on the lawn
Implementation Method 2
based on the evaluation of the spectral properties of the daylight or the sun of the light reflected on the lawn, which is measured by photodiodes
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The invention relates to an autonomous mobile platform for surface treatment, in particular lawn treatment, with a self-driving vehicle (1) designed for autonomous movement on a treatment area (6) having a boundary (7) and as a carrier for a work and/or functional unit (12) for treating the treatment area, wherein the vehicle has detection means (2, 3) for surface detection and for distinguishing it from a non-treatment surface (8) outside the boundary (7). According to the invention, the detection means comprise an electronic image acquisition unit (2) having a plurality of individual pixels, which is directed towards an image acquisition area (9) in the treatment area (6) and/or in the non-treatment area (8), and the detection means comprise a surface sensor (3) which is configured tothat its sensor detection area is in a relationship to the image detection area, preferably directed towards the image detection area (9), more preferably a sub-area (10) of the image detection area, and is linked to the image of at least one individual pixel of the image detection unit, in particular overlapping, and the detection means have classification means (11) which interact with the image acquisition unit and the surface sensor in such a way that, on the basis of a differentiation signal from the surface sensor, a plurality of individual pixels of the detection means are classified into position data of the processing area and position data of the non-processing area to generate an electronic data set of the individual pixel signals corresponding to a map depicting the boundary (7).