Insect inspiration for robot grasp and manipulation

To imitate how insects intelligently manipulate and carry objects
Description of the Project: 

Grasping and manipulation of arbitrarily shaped objects remains a difficult problem for robotics. However, foraging insects such as ants are able to make rapid and intelligent decisions about the transportability of a large variety of food items, how they should be grasped, whether to carry or drag the item, etc. Grasping by insects is largely done by the mouthparts, which resemble a standard robot gripper.

Manipulation in humans has been studied as a key capability for embodied intelligence [1] and interaction with the world, but to date very little investigation (see e.g. [2][3]) has been made of how insects - with much smaller brains - are able to carry out similar tasks. This project would involve direct study of insect behaviour, including developing methods to analyse (from video) the motor patterns and strategies exhibited in grasping decisions (this may require machine learning approaches); followed by robot modelling that reproduces the key motor and sensory systems and tests potential strategies for effectiveness in obtaining reliable and safe grasps of unknown objects. The modelling could potentially include mechanisms based on neural circuitry of the insect brain.


Resources required: 
Mobile robot base with an arm and gripper; high-speed video
Project number: 
First Supervisor: 
University of Edinburgh
Second Supervisor(s): 
First supervisor university: 
University of Edinburgh
Essential skills and knowledge: 
Familiar with robot hardware and control, programming (most likely in Python).
Desirable skills and knowledge: 
Interest in biology, computer vision (video analysis)

[1] Ritter H, Haschke R, 2015. "Hands, dexterity, and the brain."  In: Humanoid Robotics & Neuroscience: Science, Engineering & Society

[2] Cassill, D., Greco, A., Silwal, R., & Wang, X. (2007). Opposable spines facilitate fine and gross object manipulation in fire ants. Naturwissenschaften, 94(4), 326-332.

[3] Bernadou, A., Espadaler, X., Dos-Reis, V., & Fourcassié, V. (2011). Effect of substrate roughness on load selection in the seed-harvester ant Messor barbarus L.(Hymenoptera, Formicidae). Behavioral Ecology and Sociobiology, 65(9), 1763.

[4] Webb, Barbara. "Robots in invertebrate neuroscience." Nature 417.6886 (2002): 359.