1 of 36

Granular Media Characterization and Performance of a Crab-like Robot in a Variety of Environments

Emma Ries

1

2 of 36

Problem

  • Joined lab in June 2020
  • Lab focused on robot interaction with granular media, mainly play sand
  • Lack of different types of granular media

​

2

3 of 36

To determine if play sand is an adequate testing media for in-lab situations compared to real sand samples from beaches.

Objective

3

4 of 36

Revised Deliverables

  • Design of grip force tester
  • Design of dactyls to be tested via grip test Modified a pre-existing dactyl design [1] to be able to compare to previous in-lab results. [2]
  • Granular media characterization process
    • Density, Water %, Compressibility, Expansion (rolling test and shake test) Decided to be out of scope, only collected water %.
  • Test results for performance at different beaches

4

5 of 36

Final Gantt Chart

5

6 of 36

Design of the Gripper

6

7 of 36

7

8 of 36

8

9 of 36

Parts List

9

Name

Price

Quantity

Total Price

50 inch Aluminum Tripod

$14.49

1

$14.49

1 kg Load Cell

$68.00

1

$68.00

HATCHBOX PLA 1.75 mm Blue

$19.99

1

$19.99

Motor

$36.95

1

$36.95

Motor Driver

$8.49

1

$8.49

ELEGOO Uno

$12.98

2

$25.96

2000 mAh Battery

$11.99

1

$11.99

Battery Charger

$12.90

1

$12.90

Battery Plugs

$5.99

1

$5.99

Load Cell Amplifier

$15.71

1

$15.71

Jumper Cables

$6.98

1

$6.98

Arduino Assorted Components

$7.86

1

$7.86

Twine (already had)

$0.00

1

$0.00

Pavestone Natural Play Sand

$3.48

2

$6.96

Collective Price

$242.27

10 of 36

10

Arduino

Load Cell

Motor

11 of 36

Purchased Parts

11

Uno Board

Motor

HX711

Motor Driver

12 of 36

Circuit Diagram

12

13 of 36

13

14 of 36

14

15 of 36

15

16 of 36

16

17 of 36

17

18 of 36

18

19 of 36

Data Collection

19

20 of 36

MATLAB Integration with Arduino and Load Cell

  1. Tare and Calibrate Load Cell using code.
  2. Run main code.
  3. Collect 30 runs.
  4. Use modified code to produce graphs.
  5. Graph to my heart’s content!

20

21 of 36

Beach Testing

21

22 of 36

22

23 of 36

Watching Sand Dry!

23

24 of 36

24

“A”

“B”

25 of 36

Water %

​

​

“A” - ~23%

“B” - ~19%

25

26 of 36

Home Testing

26

27 of 36

27

28 of 36

Data

28

29 of 36

Raw Data from Load Cell (150 graphs!)

29

30 of 36

Previous Data

I compared my data to previous lab data. [2] A modified HEXY robot with 6 of the same dactyls was tested in Pavestone Natural Play Sand. To compare the two sets of data, I used the difference between the maximum force and the weight.

30

“Small Grip”

“Medium Grip”

“Large Grip”

“C”

Dry Sand

Wet Sand

Submerged Sand

[2]

31 of 36

My Data

Beach Tests

  • Beach “A”
    • Wet
  • Beach “B”
    • Wet

Home Tests

  • Home “A”
    • Dry
  • Home “B”
    • Dry
  • Home “C”
    • Wet

31

32 of 36

Dry Sand

32

33 of 36

Wet Sand

33

34 of 36

Conclusion

More testing needed, but this gripper seems to be a good predictor for how the robot may behave in dry sand, however, this gripper may not be a good predictor for the robot behavior in certain wet sands.

34

35 of 36

Future Work

  • Further sand characterization
  • Testing in submerged sand
  • Adding weight to gripper to directly compare to the HEXY robot
  • Testing modified robot on sand tested here to compare to other performance

35

36 of 36

Acknowledgements and References

Thank you to Nicole Graf, Alexander Behr, and Kathryn Daltorio for their data used in the papers Crab-Like Hexapod Feet for Amphibious Walking in Sand and Waves, and Dactyls and Gripping Stance for Amphibious Crab-like Robots in Sandy Substrates.

[1] N. Graf, A. Behr, and K. Daltorio, “Crab-Like Hexapod Feet for Amphibious Walking in Sand and Waves,” 2019, pp. 158–170.

[2] Graf, Nicole M., Behr, Alexander M., and Daltorio, Kathryn A., “In Revision - Dactyls and Gripping Stance for Amphibious Crab-like Robots in Sandy Substrates,” Bioinspiration & Biomimetics, 2020.

​

​

36