Weeding Robot AI Plant Classification Feedback Loop
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic lawn and garden devices lack autonomous and reliable differentiation between desired plants and weeds, especially during early growth stages, requiring user intervention and lacking intuitive user interfaces for efficient decision-making and action.
Innovation Solution
A robotic device equipped with a camera sensor system, wireless link to a handheld computing device, and AI software module that compares plant images with a database to classify plants, allowing for autonomous decision-making and action, with user interface options for user involvement and training the AI, and geometric/lawn representation for context.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If the robotic device uses AI software module to automatically distinguish plants from weeds, then the autonomy and productivity are improved, but the reliability of plant classification deteriorates especially during early growth stages
Solution Approach 1:
The system implements feedback through a user interface that allows users to review AI classification decisions and provide corrections. This feedback loop enables the system to learn from user corrections and improve its classification accuracy over time, particularly for challenging early growth stage plants where initial reliability is low.
Solution Approach 2:
The system performs preliminary actions by allowing users to pre-register and photograph plants they intend to cultivate before the weeding season begins. This creates a customized reference database specific to each user's garden, enabling more accurate classification from the start and reducing false positives during early growth stages.
2Reliability
If the system requires user intervention for plant classification decisions, then the reliability is improved, but the productivity and autonomy deteriorate
Solution Approach 1:
The system applies partial user intervention by automatically handling clear-cut classification cases while only requiring user input for ambiguous or uncertain classifications. This partial action approach maintains high productivity by automating most decisions while improving reliability by seeking user input only when necessary for accurate classification.
3Ease of operation
If the user interface provides detailed plant information and options for user involvement, then the ease of operation is improved, but the device complexity increases
Solution Approach 1:
The system uses a wireless communication interface as an intermediary between the robotic device and the user's smartphone or computing device. This intermediary handles the complexity of data processing, plant database management, and user interface presentation, while keeping the robotic device itself relatively simple. The intermediary enables detailed information display and user interaction without significantly increasing the complexity of the core robotic system.
Data Source
Figure 1
Figure 2A~2B
Figure 3
AI summary
A system comprising a robotic device and a computing device. The system is associated with a plant data base, that contains physical parameters of at least one plant. Further there is a the computing device that includes a software module forming an artificial intelligence to allow to distinguish between plants that one would like to cultivate in its lawn and/or garden environment and representations if unwanted plants, thus called weed. The system further has a user interface that allows a user to interact with the system. The user interface allows to display to a user at least one representation of a plant where the software module forming an artificial intelligence requests a user decision on whether the plant is one that one would like to cultivate in its lawn and/or garden environment and representations if unwanted plants. Thereby the system takes action or not according to the decision taken by the user. In an alternative of additional embodiment the user interface allows to present to the system with information about the plant related set-up of the lawn and/or garden environment the system takes care of.