Radar Row-Center Detection for Autonomous Vineyard Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving systems face challenges in navigating between rows of vineyards with precision, especially due to unpredictable plant growth and unreliable GPS reception, which can lead to damage of the vineyard structures.
Innovation Solution
Combining preliminary knowledge of row distances with two-dimensional radar maps to identify the center of vineyard rows, using radar wave reflections to penetrate foliage and determine the alignment of plant structures, allowing for accurate estimation of medial straight lines and intermediate trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If GPS-based autonomous driving is used in large plots, then autonomous operation is achieved, but precision between rows deteriorates
Solution Approach 1:
The patent combines GPS-based autonomous driving with vision-based row detection systems. The vehicle uses both geolocation data and real-time visual processing of row structures to achieve precise navigation between rows, merging two different sensing approaches to overcome their individual limitations.
Solution Approach 2:
The vision system acts as an intermediary between the GPS navigation system and the vehicle steering system. It processes visual information about row positions and provides correction signals to adjust the vehicle trajectory, enabling precise row-to-row navigation while maintaining overall autonomous operation.
2Loss of information
If visual techniques are used to detect rows, then row identification is achieved, but penetration through foliage deteriorates
Solution Approach 1:
The vision system segments the visual scene into different components: sky, row structures, foliage, and ground. By separately analyzing the geometric patterns of row structures (poles and cables) distinct from the foliage, the system can identify row positions even when foliage blocks direct visual access to the center.
Solution Approach 2:
The system transitions from two-dimensional image analysis to three-dimensional spatial reasoning by reconstructing the 3D positions of row structures from multiple camera views or stereo vision. This allows penetration through foliage by inferring row positions from visible portions of poles and cables that extend above or below the foliage layer.
3Measurement precision
If human operators drive between rows, then precision is maintained, but automation deteriorates
Solution Approach 1:
The autonomous vehicle performs self-correction of its trajectory by continuously detecting row positions through vision systems and automatically adjusting its path. The system serves itself by making real-time navigation decisions without human intervention, maintaining precision through onboard sensing and control systems.
Solution Approach 2:
The system implements closed-loop feedback control where the vision system continuously monitors actual row positions, compares them with the planned trajectory, and automatically generates correction commands. This feedback mechanism enables the vehicle to maintain precision navigation autonomously, adapting to variations in row alignment and environmental conditions.
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 autonomous vehicles to accurately identify and follow trajectories between vineyard rows, minimizing damage and improving navigation precision even in changing environments.
Implementation Method 1
Radars are based on microwave technologies, operating at various frequencies, for example 77 - 81 GHz. The plants and the structures of the rows, for example, the concrete, metal or wood poles generate radar wave reflections with different intensities in relation to the material they are made of.
Implementation Method 2
The radar wave therefore manages to penetrate the fronds and the maximum reflection occurs where the concentration of foliage is greatest, that is, in the centre of the row
Data Source
Figure 1A~1B
Figure 2~3
Figure 4
AI summary
A method for identifying a trajectory between rows of a plantation by means of a radar interfaced with processing means of an agricultural vehicle, the method comprising the following steps in the same order: acquisition of an approximate distance (D) between two consecutive rows of the plantation, acquisition of signals by said radar, processing of said signals to obtain a two-dimensional map of points corresponding to reflections picked up by said radar, first linear interpolation to obtain a first interpolating line (IN1) on the points of greatest intensity, second windowing of an elongated area of the two-dimensional map having an axis of development approximately parallel to said first interpolating straight line and at said approximate distance (D) from said first interpolating line, second linear interpolation of a second interpolating straight line (IN2) on points of greater intensity in the windowed area, calculation of a trajectory (T) parallel and intermediate between said first and second interpolating straight line.