IIT Bombay researchers show how animals find their way home

Path taken by the homing robot and detailed view of the robot. | Photo Credit: Special Arrangement

The researchers IIT Bombay The robots are being used to understand how animals find their way back home from unfamiliar locations, a skill called homing.

“The primary goal of our research group is to understand the physics of active and living systems. We achieve this by experimenting on centimetre-sized self-driven programmable robots. In simple terms, we design these robots to mimic the dynamics of living organisms at both individual and collective levels,” said Nitin Kumar, Assistant Professor, Department of Physics, IIT Bombay.

Homing is important for many lifeforms, whether birds flying thousands of kilometres during migration or ants returning to their colonies after foraging for food. Humans have also used this ability to train homing pigeons to carry messages over long distances.

Dr Kumar’s team has now developed a robot that mimics foraging and homing behaviour. The robot is designed to move on its own and use light to find its way back home. In a new study, they report some principles of homing based on their studies with the robot.

The foraging robot is programmed to move around semi-randomly, just as animals might wander about in search of food. This type of motion is called active Brownian (AB) motion. The robot’s direction changes frequently due to rotational diffusion, a mechanism that introduces a certain level of randomness into its path. When the robot needs to return home, it goes into a different mode that does not involve randomness inputs.

The researchers shined a beam of light on the robot; the light intensity gradually changed. The robot was programmed to find its way by following this gradient of light, similar to the way some animals use the sun or other environmental light sources to find their way.

“The homing motion is similar to that of the AB model, except that the robot has to repeatedly correct its path whenever it deviates significantly from its intended homing direction, as would be expected in real organisms,” said Dr. Kumar.

In their study, the team determined how long it took the robot to return home after deviating more and more from its homeward path. They observed that the redirection rate – the frequency with which the robot adjusted its direction to return home – was related to the degree of randomness in its path. They reported an optimal redirection rate for a particular value of randomness; beyond this rate, the frequency of redirection negated the effects of randomness and ensured that the robot reached home.

According to the researchers, this suggests that animals may have evolved to reorient themselves at an optimal rate to efficiently find their way home, regardless of noise or unpredictability in their environment.

“The observation of a finite upper limit on the return time indicates that the homing speed is inherently efficient,” said Dr Kumar. “Our results demonstrated that if animals are always aware of their home direction and always correct their path whenever they deviate from the intended direction, they will definitely reach home within a certain time.”

To validate their findings, the researchers created a theoretical model that would predict how long it would take the robot to reach home based on its behaviour. Dr Kumar said the model was able to successfully explain the robot’s behaviour and also capture specific features of its route home.

He said the model highlights the importance of redirection as a strategy and shows that frequent route correction is crucial for efficient navigation.

The team also ran computer simulations in which the robot’s movements mimicked those of some animals. They matched the experimental results, reinforcing the idea that randomness and redirection work together to optimise homing. “When we applied this model to the trajectories of a real biological system of a flock of homing pigeons, it showed good agreement with our theory, validating our hypothesis of increased efficiency due to frequent course corrections,” said Dr Kumar.

It is stated that light-based cues are just one of many cues involved in real-world navigation; others include social interactions, changing landscapes, and other environmental factors. “In our future research, we aim to model these scenarios in our experiment using a combination of spatio-temporal variations in light intensity and physical obstacles,” Dr. Kumar added.