A new Traffic cop algorithm helps drones stay on task

By keeping data fresh, the system could help robots inspect buildings or search disaster zones.


Time-sensitive applications, such as search and rescue missions, smart factories, and smart city intersections, increasingly rely on collaborative multi-agent systems. Traditional communications networks are not designed for large-scale multi-agent systems that require time-sensitive information sharing to interact efficiently. WiFi is a popular alternative for deploying time-sensitive applications since it is cheap, reliable, and widely available in sensors, cameras, and other sensors.

When many robots send time-sensitive data over a wireless network, a data traffic jam might occur, making it difficult to provide a useful, real-time report. New data is critical for drones searching for survivors in a disaster zone or robots checking a building.

Engineers at MIT have developed a method for tailoring any wireless network to handle a large load of time-sensitive data from various sources. Their novel strategy, WiSwarm, configures a wireless network to control the information flow from many sources while ensuring the network relays the most recent data.

The method developed by the team allows many robots to communicate over WiFi networks without the need for expensive communications and processing hardware.

Vishrant Tripathi from MIT said, “What happens in most standard networking protocols is an approach of first come, first served. A video frame comes in, and you process it. Another comes in, you process it. But if your task is time-sensitive, such as trying to detect where a moving object is, then all the old video frames are useless. What you want is the newest video frame.”

Modiano explains, “Age-of-information is a new metric for information freshness that considers latency from the perspective of the application.” 

The team developed WiSwarm for prioritizing age-of information. This scheduling algorithm can be run on a centralized computer and paired with any wireless network to manage many data streams and prioritize the most recent data.

The team used their method to modify a conventional WiFi router, showing that the tailored network could act as an efficient traffic cop, prioritizing and relaying the most recent data to keep multiple vehicle-tracking drones on task.

The algorithm determines which source in a network should send data next, using a “last in, first out” protocol to send their freshest piece of data through the wireless network to a central processor.

WiSwarm determines which source in a network should send data next by weighing three factors: a drone’s general weight, priority (for example, a drone tracking a fast vehicle may have to update more frequently and thus would have higher priority over a drone tracking a slower vehicle); a drone’s age of information, or how long it has been since a drone has sent an update; and a drone’s channel reliability, or likelihood of successfully transmitting datum.

By multiplying these three parameters for each drone at any given time, the algorithm can schedule drones to report updates through a wireless network one at a time without clogging the system. 

The team tested out their algorithm with multiple mobility-tracking drones. The researchers outfitted flying drones with a small camera and a basic Wi-Fi-enabled computer chip, which they used to relay images to a central computer continuously.

When pairing the network with their algorithm, the computer was able to receive the freshest images from the most relevant drones, which it used to send commands then back to the drones to keep them on the vehicle’s track. 

When the researchers tested the system with two drones, they discovered that it could relay data that was two times fresher, resulting in six times better tracking. As the system was expanded to five drones and five ground vehicles, WiFi alone was unable to handle the increased data traffic, and the drones quickly lost track of the ground vehicles. WiSwarm improved the network’s capabilities, allowing all drones to continue tracking their cars.

Tal said, “Ours is the first work to show that age-of-information can work for real robotics applications.” 

Karaman said, “Imagine self-driving cars come to an intersection that has a sensor that sees something around the corner. Which car should get that data first? It’s a problem where timing and freshness of data matters.”

In the near future, cheap and nimble drones could work together and communicate via wireless networks to survey buildings, agricultural fields, and wind and solar farms. The concept could be critical for controlling data streaming throughout smart cities.

The result shows that WiSwarm provides far better tracking than WiFi with only two UAVs.

Journal Reference:

  1. Vishrant Tripathi, Igor Kadota, et al. WiSwarm: Age-of-Information-based Wireless Networking for Collaborative Teams of UAVs. arXiv DOI: 10.48550/arXiv.2212.03298