4️⃣AGV Forklift Bin Stacking Solution Introduction

Solution Overview

Core advantages

1. High-precision recognition: According to the deviation value recognized by the camera, the forklift can flexibly adjust the stacking posture, the positioning accuracy is up to ±5mm, the angle accuracy is up to 0.1 °, and the recognition range of the cage offset is ±15°;

2. Intelligent cage recognition: Self-developed AI algorithm directly outputs real-time relative position information of loading and unloading bins to adjust forklift attitude in real-time;

3. Real-time data processing: The data processing time per frame is less than 100 milliseconds, ensuring the efficient operation of forklift operations;

4. Multi-layer stacking: support 1-4 layer stacking, 2 * 2 stacking mode

Animated demo

Camera selection advice

Camera installation method

• Installation method 1: Single camera solution
• Position: The camera is fixed on the lifting mechanism, shooting horizontally, and responsible for the identification of the forklift arm loading bin on both sides.
• Height requirement: The camera is 35-55cm away from the lower edge of the forklift arm to ensure that the position error of the lifting mechanism does not exceed 1cm.
• Features: The single camera is responsible for identifying the legs of the bin on both sides of the forklift arm.

• Installation method 2: Dual camera solution
• Position: The camera is fixedly installed on both sides of the forklift, 30-40cm away from the surface of the forklift arm, and the shooting angle is 45 ° with the ground.
• Features: The left camera shoots the left forklift arm, and the right camera shoots the right side, respectively recognizing the two sides of the bin to achieve accurate stacking.

Communication method

• Request information: send a request to obtain the X, and Y coordinates and angle (θ) of the bin.

• Return information :
• X, Y coordinates: in millimeters, 32-bit floating-point type.
• Theta Angle: Range from -90 ° to 90 ° for precise identification of deviations.
• Error Code: recognition status (0: success, -1: camera exception, -2: positioning failed, -3: algorithm error).

Specific workflow

1. Positioning: The forklift moves 2.4-2 meters from the front of the bin.

2. Camera recognition: The forklift arm is raised to the set position, triggering the 3D camera to recognize the posture of the basket.

3. Adjust position: The forklift adjusts the angle and position according to the deviation information output by the camera.

4. Pick-up operation: When the forklift confirms the correct position of the bin, the forklift arm performs a pick-up operation.

1. Pre-positioning: The forklift is moved to about 80cm in front of the stacking bin, and the forklift arm is raised to 50cm from the upper surface of the bin.

2. Camera recognition: The system fine-tunes the left and right angles by outputting the deviation value of the 3D camera.

3. Precise alignment: After confirming that the angle and position meet the threshold, the forklift is aligned through fine-tuning.

4. Complete stacking: After confirming the accuracy, the forklift arm descends to complete the stacking operation.

Notes on Deployment

1. Environmental requirements: avoid serious dust and humid environment, so as not to affect camera performance.

2. Installation requirements: Ensure the rigidity and stability of the 3D camera bracket to avoid affecting recognition accuracy.

3. Regular maintenance: Regularly clean the camera lens to prevent dust coverage and ensure clear camera vision.

Recognition effect display

Application case sharing