Robotic and uncrewed submersibles could operate more effectively by mimicking the way some fish probe their surroundings with electric fields, say researchers.
Many marine and freshwater fish can sense electric fields, but some also generate their own weak fields over short ranges to help navigate, identify objects, and even communicate with other fish.
Malcolm MacIver and colleagues at Northwestern University in Chicago, US, studying the biomechanics of these "weakly electric fish", have now come up with an artificial electric-field sensing system. They say it could ultimately give robot submersibles the same additional sensory capabilities.
"Currently, no vehicle is manoeuvrable enough to do work in tight quarters, such as coral reef monitoring, underwater structural inspection, or searching a submerged vessel," MacIver told New Scientist. "To do so requires not only a high amount of agility, but also being able to sense in all directions, so that you do not collide with nearby obstacles. Electro-location is perfect for this."
The researcher’s electro-location system consists of two field-emitting electrodes and two voltage-sensing electrodes. These electrode pairs are arranged at opposite corners of a diamond, and were submerged in shallow water for testing purposes.
If the electric field is not disturbed by anything in the water, the two pairs of sensor electrodes should provide identical readings. But if something is placed within a few centimetres of the sensors, the field is disturbed and causes a variance in the sensor readings.
Since the position of the object affects the way the electric field is disturbed, a computer can use a series of sensor readings to determine the object’s location. It is even possible to detect small objects by amplifying subtle perturbations in the field.
To test the set-up, the four sensors were attached to a robot arm and lowered into a shallow tank of water. The sensors were then moved around a sphere fixed to the bottom of the tank while readings were taken.
The experiment was repeated using spheres of two different sizes, made of both metal and plastic, and in both fresh and salty water. In all cases it was possible to pinpoint the position of the object to within 3 millimetres after 10 movements.
MacIver believes compact, agile submersibles could one day be covered in many electro-sensors – although he notes that weakly electric fish use many thousands. So the next step, he believes, is to develop an array of sensor electrodes and attempt more complicated tests.
The team also hopes to be able to gather data on the size, shape and composition of objects, as these properties also affect how it disturbs an electric field.
Steve McPhail, who designs autonomous underwater vehicles at the UK’s National Oceanographic Centre, part of the University of Southampton, also sees potential in the idea. "It sounds like this would be useful for small, agile subs operating close to the seabed or in tight spaces," he says. "An obvious advantage is that the sensors are quite cheap."
MacIver admits that it will be a long while before electric field sensors are anywhere near as sophisticated as those found in nature. "It has taken the machine vision community many years to ‘teach’ computers how to perceive simple objects using light," he notes. "We are just starting the process of understanding how to perceive simple objects using ‘electro-sense’."
You can see lots of video clips and animations relating to the work of the MacIver group on weakly electric fish on the group’s homepage.
The research was presented at the International Conference on Robotics and Automation (ICRA), held in Rome, Italy, in April 2007.
Via New Scientist