Evaluation Setup
Physical grasping experiments with objects of varying size, shape, and weight.
Benchmarks:
- Size Variation Test (Grasp objects from 5.6mm to 115mm diameter) [New]
- Load Capacity Test (Lift weights up to failure) [New]
- Dynamic Grasping (Catch moving/dynamic objects via whipping) [New]
Metrics:
- Success rate
- Payload-to-weight ratio
- Graspable object diameter range
- Statistical methodology: Not explicitly reported in the paper
Key Results
| Benchmark |
Metric |
Baseline |
This Paper |
Δ |
| Load Test |
Payload-to-Weight Ratio |
0.0384 |
10.0 |
9.9616
|
| Size Variation Test |
Object Diameter |
60 |
5.6 |
-54.4
|
| Automatic Grasping |
Success Rate |
Not reported in the paper |
94.4 |
Not reported in the paper
|
Main Takeaways
- Logarithmic spiral geometry naturally provides a wider range of graspable sizes compared to constant curvature designs.
- Smaller taper angles increase the total workspace, while larger taper angles increase load capacity and ability to grasp smaller objects.
- The 'climbing' strategy (uncurling along object) is robust to position uncertainty and allows operation in confined spaces.
- Scalability is demonstrated effectively: 1cm micro-robot (ant manipulation) to 1m macro-robot (drone-mounted).