This project implements a comprehensive pick-and-place system for a robotic arm, designed to handle both static and dynamic blocks in simulated and real-world environments. The system is capable of detecting, grasping, and stacking blocks onto a designated platform under both stationary and motion-based conditions.
- Handles static and dynamic block manipulation
- Adapts to team-specific configurations (red or blue)
- Utilizes forward and inverse kinematics for precise movements
- Implements real-time detection of AprilTags for pose estimation
- Employs dynamic motion prediction for blocks on a spinning table
- Optimizes performance through joint-angle caching and safe movement planning
System Overview
The system uses a set of criteria for dynamic block picking, as illustrated in the image above. This ensures accurate and reliable grasping of moving blocks on the spinning table.
The system demonstrates high accuracy in handling both static and dynamic blocks:
Static Blocks Figure 1: 4 Static blocks stacked
Dynamic Blocks Figure 2: 4 Dynamic Blocks
Combined Stack Figure 3: 6 Stacked blocks (combined static and dynamic)
| Parameter | Value (Average) |
|---|---|
| Number of static blocks picked | 4.0 |
| Number of dynamic blocks picked | 3.5 |
| Total time for algorithm execution (seconds) | 455.5 |
| Average height of static block stack (cm) | 20.0 |
| Average height of dynamic block stack (cm) | 15 |
| Combined average height of both stacks (cm) | 25 |
The system is built around the PickAndPlace class, which encapsulates all functionalities for organized and maintainable code. Key components include:
- Class initialization with team color and hardware interfaces
- World and base frame configuration
- End-effector pose definitions for safe movements
- Block detection and validation methods
- Grasping strategies for static and dynamic blocks
- Tower building and block placement logic
- Angular Velocity Discrepancy: Addressed by implementing dynamic estimation of the spinning table's speed.
- Gripper Timing and Force Feedback: Compensated with alternative sensing and control strategies.
- Environmental Variability: Handled through robust image processing and adaptive sensing techniques.
The system incorporates comprehensive logging for troubleshooting and performance analysis:
Logs
Figure 4: System logs for performance analysis
- Implement more advanced path planning algorithms (e.g., RRT)
- Enhance error handling and recovery mechanisms
- Integrate machine learning for improved dynamic object tracking
- Develop alternative feedback mechanisms to compensate for unreliable force feedback
This pick-and-place system demonstrates robust performance in handling both static and dynamic blocks, showcasing the potential for advanced robotic manipulation in various industrial applications. The project highlights the importance of adaptive strategies and real-time adjustments in robotic systems operating in dynamic environments.