GIOVANNI PINI
Senior Robotics Software Engineer
HOPE Technik, Singapore
 

RESEARCH OVERVIEW

My research focuses on a branch of robotics named swarm robotics. Swarm robotics draws inspiration from social insects in the implementation of large scale multi-robot systems.

Social insects, such as bees or ants, display an important characteristic: the colony can produce outcomes that far exceed the capabilities of the single individual. Ants or bees are very simple organisms and their capabilities are limited, nevertheless we can observe how sophisticated a bee hive or an ant nest is, or how efficiently those animals can find sources of food and organize their work. And the best part of the story is: none of the ants or the bees has the role of organizing the activities. Each individual performs its own actions following simple rules and using local information only.

In social insects, complex outcomes at the colony level arise from the interaction within individuals and between the individuals and their environment. The fact that each individual acts on its own following simple rules makes the system scalable, robust, and tolerant to faults. Ants and bees are interested in those properties as they allow them to survive, roboticists are interested in obtaining the same properties in their systems.

The ultimate goal of swarm robotics is to understand the mechanisms social insects use to organize themselves in a distributed fashion, and to apply them to robotic systems. In this way, one can employ a system composed of very simple and cheap robots, and employ it in situations where a single robot is not enough. Following the principles of simplicity and decentralization, one can usually build systems that are scalable, robust, and tolerant to faults.

Tolerance to faults and cheapness make swarm robotics appealing for applications where a high failure rate at the level of the single robot can be expected, such as in space exploration or other tasks performed in harsh conditions or dangerous environments. Another important field of application is for all those tasks that require miniaturization of the robots.

Swarm robotics is a vast field, different aspects are being studied. My personal research work focuses on three main topics: task allocation, task partitioning, and evolutionary robotics. My work in each of these topic is discussed and presented in a dedicated section.