New method developed by ARL helps robots stay networked
The robot maps the environment by selecting a frontier to explore next. A frontier selection algorithm uses inverse distance from the robot's current position, the RSS gradient, and the average RSS trend. The figure shows how the robot selects frontier 1, corresponding to an increase in RSS.
|
|
Story Highlights
- ARL, CISD scientists recently developed a new method that allows a team of robots to maintain radio connectivity.
- Robots use analysis of local radio signal strength to plan team moves.
- Robots use signal-to-noise ratio.
U.S. Army Research Laboratory (ARL) scientists have recently developed a new method that allows a team of robots to maintain radio connectivity when executing a complex mission such as collectively mapping the interior of a building.
There are many challenges that exist in the area of mobile robotics, including maintaining radio-based connectivity in a complex environment, which is hindered by the rapid fading or fluctuation of received radio signal strength (RSS).
Jeffrey Twigg, Dr. Jonathan Fink, Dr. Paul Yu, and Dr. Brian Sadler of ARL's Computational and Information Sciences Directorate (CISD) have developed an algorithm that fulfills their goal of enabling robots to identify a radio source and map the connectivity region in an unknown, indoor environment.
The algorithm developed pairs geometric exploration of an unknown environment with a radio-source seeking behavior which drives a robot towards areas consisting of increased signal-to-noise ratio.
The robots are able to repeatedly carry out inexpensive analysis of local radio signal strength and use the information to plan their team moves in order for them to all remain networked.
The robot used in the experiments for this method, which is able to record RSS indicator, is known as the Scarab, which is a small indoor ground platform equipped with a differential drive system, onboard computation, wireless communication, a Zigbee radio which is used for experimental measurements, and a scanning laser range finder used to provide self-localization.
With this development, it is possible for robots to be guided to a signal source in an efficient manner, avoiding both random and exhaustive exploration and overwhelming the computing abilities of the small robots.
Also, due to the fact that this newly developed method does not require previous knowledge of the environment being explored, with the exception of the challenges of fading, and does not require a map prior to the start of the exploration, it can be used in a wide array of scenarios.
"Maintaining network connectivity in a complex and unknown environment is a challenge that the Army has worked on for decades, and we are working on fundamental techniques that employ autonomous agents to maintain connectivity, and continuously provide critical situational awareness services to the Soldier," said Sadler.
Sadler also said that he and the other scientists apply the same ideas to collaborating robots that can explore an unknown area, make a map, and provide sensing such as finding biological or other threats.
"The problem is richly multi-disciplinary, combining network science, sensing and autonomy," Sadler stated.
This new method is a significant advancement in the areas of networking and robotics, and has created great leverage for the development of effective teams of small robots for Army missions.