Engineers at the University of California, San Diego have developed low-cost, low-power technology that helps robots map their indoor environments. The system helps robots navigate even in low light or when there are no recognizable landmarks or features.
The team of researchers belongs to the Wireless Communications Sensing and Networking Group, led by UC San Diego Electrical and Computer Engineering Professor Dinesh Bharadia. It will be showcased at the International Conference on Robotics and Automation (ICRA) 2022 in Philadelphia, which runs May 23-27.
The research has been published in IEEE Robotics and Automation.
A whole new approach
The newly developed technology has sensors that rely on WiFi signals to allow the robot to map its environment and path. The system is a whole new approach to navigating indoor robots, and it’s unique from previous ones that use optical light sensors like cameras and LiDARs.
“WiFi” sensors use radio frequency signals instead of light or visual signals to see, allowing them to work better in environments where cameras and LiDARs have trouble. These types of environments are usually dimly lit, changing, and repetitive like long hallways.
Alternative to LiDARs
WiFi is helping the technology achieve its status as a cost-effective alternative to LiDARs, which are expensive and power-intensive.
“We are surrounded by wireless signals almost everywhere we go. The beauty of this work is that we can use these daily signals to do indoor location and mapping with robots,” Bharadia said.
Aditya Arun holds a PhD in Electrical and Computer Engineering. student in Bharadia’s lab and first author of the study.
The researchers built the prototype system with off-the-shelf hardware. It consists of a robot equipped with WiFi sensors built from commercially available WiFi transceivers. These WiFi sensors transmit and receive wireless signals to and from WiFi access points in the environment, and this communication is what allows the robot to map its location and direction of movement.
Roshan Ayyalasomayajula also holds a PhD in Electrical and Computer Engineering. student in Bharadia’s lab, as well as a co-author of the study.
“This two-way communication already happens between mobile devices like your phone and WiFi hotspots all the time – it just doesn’t tell you where you are,” Ayyalasomayajula said. “Our technology relies on this communication to perform localization and mapping in an unfamiliar environment.”
WiFi sensors first ignore the location of the robot and where WiFi access points are in the environment. As the robot moves, the sensors call the access points and listen to their responses, which are then used as landmarks.
Each incoming and outgoing wireless signal carries its own unique physical information that can be used to identify where robots and access points are in relation to each other. Algorithms allow WiFi sensors to extract this information and perform these calculations. The sensors continue to gather more information and can eventually pinpoint the direction of the robot.
The technology was tested on one floor of an office building, where multiple access points were placed around the space. A robot was then fitted with the WiFi sensors, along with a camera and LiDAR to perform measurements for comparison. The team controlled the robot and ran it around the floor several times. He also turned corners and walked down long, narrow hallways with brightly and dimly lit spaces.
Tests have shown that the location and mapping accuracy provided by WiFi sensors is comparable to that of commercial camera and LiDar sensors.
“We can use WiFi signals, which are essentially free, to perform robust and reliable sensing in visually challenging environments,” Arun said. “Wi-Fi sensing could potentially replace expensive LiDARs and complement other low-cost sensors such as cameras in these scenarios.”
The team will now work to combine WiFi sensors and cameras to develop even more comprehensive mapping technology.
#WiFi #helps #robots #navigate #indoor #environments