Non-Stop Product Sorting Automation using Robostar RA007 and Deep Learning

Industry > LogisticsApplication > ClassificationSector > Manufacturing

This solution utilizes 3D vision, deep learning algorithms (Instance Segmentation), reinforcement learning, and motion planning (OMPL) algorithms to automate tasks such as product sorting, assembly, and packing.

By using cost-effective 3D vision to recognize product types, product locations, and box positions in real-time, the system can automatically compensate for and continue operations even if the positions of products or boxes change.

Furthermore, with two vision sensors and high-speed processing, it can significantly reduce cycle times—the most critical factor in logistics sites.

Since the system compensates through learning even if the conditions change due to misalignment of the vision, robot, or work jigs, it ensures easy operation and maintenance while minimizing downtime.

Components

Robot	Robot: Robostar 6-Axis Articulated RA007 Payload: 7kg, Reach: 930mm, Repeatability: ±0.03mm, Weight: 38kg Gripper: Vacuum gripper (Custom-made) 3D Vision: 2 units of 3D stereo cameras (equipped with proprietary algorithms)

Workflow

STEP 1.	Separately recognize each product from an input box containing a mix of various items.
STEP 2.	Pick products starting from those stacked on top.
STEP 3.	Load the picked products into designated positions within the discharge box.
STEP 4.	Continue operations until all items in the input box are sorted into individual boxes. (The system recognizes Pick & Place locations via vision and performs tasks regardless of changes in the positions of input or discharge boxes.)

※ The content provided by Bigwave Robotics is protected under U.S. and Canadian copyright and intellectual property laws. Unauthorized reproduction, distribution, or use of this content is strictly prohibited.

Features

Compact and Simple Installation

Easy Robot Teaching: Teaching time is reduced through automated vision calibration and teaching using deep learning.

No Recalibration Needed: Even if work conditions change, manual recalibration and re-teaching are unnecessary.

Superior Recognition: Deep learning-based 3D object recognition provides excellent detection capabilities for a wide variety of items.

Autonomous Learning: Automatically learns the optimal picking position according to the product's shape.

Compact and Simple Setup

No Fixed Positioning Required: Since input and discharge boxes do not need to be fixed in specific locations, there is no need to manufacture separate feeders.

Seamless Integration: The robot can be installed directly in the existing workspace previously used by human workers.

Fast Return on Investment (ROI)

Rapid Cycle Times: By using high-speed industrial robots and independent standing vision systems, cycle times can be shortened by more than 2 seconds compared to Hand-Eye solutions.

Operational Optimization & Maintenance Cost Reduction: Shortens cycle times by continuously learning and optimizing work postures and paths.

Automatic Compensation: Even if the layout shifts or the robot becomes misaligned due to aging, the system automatically compensates and continues tasks.

Stable Productivity: Minimizes downtime to ensure consistent production levels.

Results

ROI Period	null

Please log in to view