Army researchers play 'Marco Polo' to locate personnel, robots
- By Patrick Marshall
- Nov 15, 2018
GPS has become the go-to technology for consumer navigation, whether it’s a pedestrian looking for a coffee shop or a driver seeking an address in an unknown part of town. Unfortunately, since GPS signals can be spoofed, can't penetrate buildings and their satellites can be destroyed, the military can’t count on using GPS to locate military personnel and robots deployed in the field.
Like radar and sonar, broadcast wireless signals can be used to locate objects by measuring the time it takes for a signal to bounce back to a receiver. But that technology has limited success in object-rich environments, such as the inside of an office building or manufacturing plant. In a complex environment it may be impossible to determine where a received wireless signal originated.
“In the presence of obstacles, the wireless signal gets blocked entirely, or scattered in different directions,” Army Research Laboratory Researchers Fikadu Dagefu and Gunjan Verma told GCN. “Think of it like shooting a basketball," they wrote in an email describing the technology. If no one is guarding the shooter, it is easier to get a shot off. "But in the presence of obstacles (big tall defenders), the ball is likely to get blocked or deflected. For wireless signals, buildings and other obstacles alter the direction of the signal.”
Dagefu and Verma, along with other on the ARL team, have developed an algorithm for analyzing received wireless signals to determine their direction of arrival. The trick, according to the researchers, is to continuously measure the strength of wireless signals rather than just the direction from which they seem to be coming.
Imagine the game children play in a swimming pool -- Marco Polo – where one child who has his eyes closed calls out "Marco" and tries to locate the other players by listening to their responses of "Polo."
"We can move closer to the target child, even in the absence of vision, by judging whether or not the child’s voice is getting louder,” they said. It’s the same with a wireless signal. “If we move towards the source, it gets stronger.”
And by measuring how quickly the signal grows stronger, a receiver can tell how directly it is approaching the source. “Teasing out this direction information in the presence of noise and potential echoes that could be created due to nearby obstacles in a theoretically grounded manner is what the proposed algorithm does,” the researchers said.
The research team has tested the approach using publicly available datasets and in-house simulations in the 40MHz and 2.4GHz bands, as well as data from high-fidelity simulations. According to ARL, the technique works well under conditions “in which classical phase or time of arrival based estimates would fail.”
“In addition to not requiring any fixed infrastructure, the proposed technique also does not rely on any prior training data, knowledge about the environment, multiple antennas, or prior calibration between nodes,” the team reported.
Patrick Marshall is a freelance technology writer for GCN.