Crop-Row Robot Navigation Using Grid-Based Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous navigation of agricultural robots between two rows of plants is challenging due to varying environmental conditions and the uncertainty of GPS signals in orchards and groves, leading to potential collisions.
Innovation Solution
A method and system utilizing two sensing devices, sensor A and sensor B, which detect electromagnetic or sound waves with different frequencies and fields-of-view, to create a two-dimensional grid for data point collection and fusion, allowing the robot to autonomously navigate by calculating fusion function values and classifying activation levels of grid cells to determine the centerline path.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS-based autonomous guidance is used for agricultural equipment, then navigation uniformity is improved, but GPS signal visibility becomes uncertain and multipath errors occur in dense canopies
Solution Approach 1:
The patent introduces an intermediary system consisting of overhead sensors (lidar, cameras, GPS receivers) mounted on fixed structures that mediate between the satellite GPS signals and the autonomous vehicle. This intermediary captures images and GPS data from elevated positions, providing reliable navigation information that penetrates through dense canopies where ground-level GPS signals are blocked or corrupted by multipath effects.
Solution Approach 2:
The patent transitions from ground-level GPS reception to overhead/aerial dimension for signal acquisition. By mounting sensors on overhead structures (towers, poles, or other elevated positions), the system accesses GPS signals and visual data from a different spatial dimension that is not obstructed by the canopy, thereby resolving the signal blockage problem while maintaining navigation uniformity.
2Productivity
If autonomous navigation is implemented between plant rows, then productivity is improved, but collisions with plants occur due to varying environmental conditions
Solution Approach 1:
The patent employs preliminary action by continuously capturing images and GPS data overhead before the autonomous vehicle reaches potential obstacle zones. The system pre-processes this data to create a forward-looking view of the plant rows, enabling the vehicle to anticipate and adjust its path in advance, thereby preventing collisions rather than reacting after contact occurs.
Solution Approach 2:
The patent implements a feedback loop where overhead sensors continuously monitor the environment, the processor compares actual plant row positions with the planned centerline path, and the steering system makes real-time corrections. This closed-loop feedback enables the vehicle to adapt to varying environmental conditions (wind, soil conditions, terrain) while maintaining accurate navigation between rows.
3Measurement precision
If sensor fusion with multiple sensing devices is used, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent applies universality by designing the overhead sensor system to perform multiple functions simultaneously: GPS signal reception, visual imaging of plant rows, and environmental monitoring. By consolidating these functions into a single overhead platform, the system achieves high measurement precision through sensor fusion while minimizing the increase in overall device complexity compared to having separate ground-level systems for each function.
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
Enables reliable and collision-free autonomous navigation of agricultural robots between rows of plants under changing environmental conditions by accurately determining the centerline path through sensor fusion and grid-based data processing.
Implementation Method 1
each sensing device is a device which detects electromagnetic waves or detects sound waves
Implementation Method 2
each sensing device is a device which detects electromagnetic waves or detects sound waves
Implementation Method 3
calculating a fusion function value, Γ(h,v), for each cell (h,v) of said two-dimensional grid
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method, system and robot for autonomous navigation thereof between two rows of plants, wherein said robot comprises two or more sensing devices, sensor A and sensor B, mounted thereon and moves forward along an axis parallel to the rows of plants, being autonomously steered by exerting angular corrections to place the robot as close as possible to the centerline between the rows of plants, wherein each sensing device is a device which detects electromagnetic waves or detects sound waves, and wherein said method and said system comprises the following: (i) defining a two-dimensional grid of square cells in said plane; (ii) dividing the two-dimensional grid of square cells into IG•JG groups of cells; (iii) obtaining a maximum of k data points using sensor A and a maximum of m data points using sensor B; (iv) converting each data point into a discretized data point; (v) calculating a fusion function value for each cell (h,v) of said two-dimensional grid; (vi) calculating for each group of cells (i,j) (omij): (a) the cumulative total of fusion function values, CUMij(h); and (b) the sum of all cumulative totals of fusion function values, SUMij; wherein when: - SUMij ≥ ||0.4•OM|| said group of cells is classified as high-activated; and - SUMij ≥ ||0.2•OM|| and < ||0.4•OM|| said sub-volume is classified as low-activated, wherein OM is the maximum SUMij value determined for any group of cells (i,j) (omij); (vii) moving the robot: (a) by turning right; or (b) by turning left; or (c) forward without turning, depending on whether each group of cells is low-activated, high-activated or not activated.