Search:  

 Blog     e-Newsletter       Resource Library      Directories      Webinars     Apps
AFCEA logo
 

Small Machines Weave Communications Web

November 2008
By Henry S. Kenyon

 
The Defense Advanced Research Projects Agency’s (DARPA’s) LANdroids program seeks to solve urban communications problems faced by warfighters. The program is creating intelligent autonomous robotic relay nodes that can be scattered by soldiers as they enter a building.
Smart, mobile nodes will configure, manage tactical voice, data links.

Sometime soon, swarms of autonomous robots may help battlefield communications networks stay up and running even in the most challenging battlefield environments. Each individual machine is a mobile communications node. When grouped together, these smart relays will automatically form a network and realign themselves to maintain links in the face of jamming, radio interference or complex, radio-unfriendly terrain such as buildings.

The goal of the U.S. Defense Advanced Research Projects Agency’s (DARPA’s) LANdroids program is to create small, inexpensive robotic relay nodes that dismounted soldiers can drop as they operate in urban areas. The smart nodes then reconfigure to form a mesh network to temporarily link communications in an area. According to DARPA, the robots also will perform self-healing network realignment if individual nodes are destroyed.

A key motivator for the program is maintaining communications links in urban environments. Dr. Mark McClure, DARPA’s LANdroids program manager, Arlington, Virginia, notes that communications are unreliable in urban settings. This uncertainty means that warfighters often lose contact with each other and have difficulty controlling deployed sensors or robots. “Anything that relies on communications may not work properly. The root cause is basic physics. Obstacles like buildings, walls, even cars, block radio waves and hinder communications,” he explains.

Two key firms contributing to the program are Lockheed Martin and iRobot Corporation. Lockheed Martin is working with DARPA on the wireless and software aspects of LANdroids. Peter Drewes, Lockheed Martin’s LANdroids program manager, Cherry Hill, New Jersey, notes that in the effort’s first phase, the company is working with commercially available robotics and radio equipment. The firm is utilizing low-cost commercial gear because DARPA stipulated that the deployed robots be as inexpensive as possible.

Autonomy is a key challenge for the program. McClure explains that one of the technical hurdles facing the effort is creating a distributed intelligent control capability that enables individual robots to decide when and where to move to optimize the mesh network’s performance. The LANdroids perform this function by communicating with their neighbors and exchanging network information.

“How do you take the information that each node has and share it so that there’s an adequate common picture, without interfering with the primary goal of allowing soldiers to use this as a communications node?” Drewes posits.

Lockheed Martin’s engineers are contributing their experience in designing, developing and supporting self-forming networks and data collaboration through minimal communications. In its program requirements, DARPA indicated that the LANdroid robots had to balance mobility and signal strength to preserve their battery power. The company’s researchers are working on the core software that will manage network configuration and control power consumption.

Researchers are studying the best means to use the mobile nodes. Mobility remains a limiting factor for the small robots. Lockheed Martin engineers are developing software solutions to move around the mobility restrictions. Drewes explains that the nodes are placed to maintain what he refers to as “good enough” communications. This requirement permits troops to open up a communications link in a building or to strengthen existing networks. “We’re looking for a globally good enough solution to send communications through, which may not be quite optimal for each individual node,” he contends.

Drewes notes that communications in and around elevator shafts are another example of the difficulty of maintaining links indoors. Warfighters encounter radio frequency dead zones in buildings caused by walls, debris and furniture. Because buildings are multifaceted challenges, the individual LANdroid nodes can optimize their networks by moving closer to each other. Under ideal conditions, the nodes have a range of roughly 100 meters, but in a modern office building, that range may only be six or seven meters. “It’s completely environmentally specific in terms of the potential range between two nodes,” he says.

One of the LANdroids program’s objectives is to support a fully autonomous network with no human intervention after the robots are tossed or dropped by warfighters. “The goal is that there is no human optimization or self-optimization. That way, you’re not losing any manpower in an operation, just like you wouldn’t want a soldier to back up 20 meters just to send a message,” he relates. The primary reason to have a mobile communications capability is to avoid these types of situations, he says.

The LANdroids program was launched in early 2008. Drewes notes that the program’s feasibility already has been proven through static tests. The first-phase goal will have small groups of five to 15 LANdroids autonomously set up and manage, self-configure and self-optimize a mesh network. He adds that this also would include finding communications links if a node falls out of the network.

DARPA plans to leverage the first-phase research as quickly as possible and then build a more capable platform and optimize the software. Testing and field trials would follow this modification phase.

Drewes claims that the core LANdroid system technology is operational. Current testing is focusing on components and software robustness as the system becomes more complex. But he maintains that the system already is fully autonomous. At the end of the first phase, several scheduled experiments will provide DARPA with metrics.

 
 
Massachusetts-based iRobot Corporation is leveraging its many years of experience in designing and manufacturing battlefield robots, such as these systems developed for the U.S. Army’s Future Combat Systems, to create small mobile nodes for DARPA’s LANdroids program. The effort will develop tiny, rugged machines that can autonomously maintain and manage communications links.
Managing networks via software is not particularly difficult, Drewes relates. The challenge is one of perspective. He uses the analogy that LANdroids operate like playing a game of “Marco Polo” in office cubicles. The machines essentially are using their “ears” to sense the environment. Program engineers are working on the software to enable the robots to work in the physical domain, while understanding what is being received in radio frequency. “That’s the hardest thing with the continual translations of different information. You see where it is. You see what’s going on; it [the robot] doesn’t see that,” he says.

In the event of jamming or loss of communications between one node and the network, the robots will move together to solve the problem. Drewes indicates that by decreasing the distance between each other, the machines can greatly minimize the effects of jamming or radio interference.

Designers also are considering the question of power use. Drewes explains that engineers are trying to optimize the machines to operate at specific signal strengths, depending on the circumstances, to preserve energy. He notes that the crux problem is whether it makes more sense for the robot to increase its transmission power or to move closer to another node. Each option uses battery power. “If you spend all day moving, you might find an optimal communications area, but there is a cost to movement that has to be balanced out against the costs to transmit, from an energy perspective,” he says.

When it becomes necessary for the nodes to sense their immediate physical surroundings, Drewes asserts that sophisticated sensors may not be necessary. The robots navigate mostly by radio transmissions, with some physical input from platform-mounted sensors. “Basically, we’re bumping our way through the environment and using that to perceive approximately how far we can go before we hit anything. What can we infer about what we hit? How does our RF [radio frequency] signal change when we hit it?” he remarks.

iRobot is responsible for developing the robotic platform. Christopher Jones, iRobot’s LANdroids research program manager, Bedford, Massachusetts, notes that the program seeks to make small, lightweight, inexpensive robots that can fit comfortably in a soldier’s hand. Ideally, soldiers will be able to carry several of the robots. He indicates that the robot nodes must be tough enough to survive being tossed into a room or hallway.

Although the smart robots can autonomously navigate to a new position, he asserts that they are not intended to travel more than several meters once they have been deployed. To maintain immediate situational awareness, Jones says that the LANdroid robots will use a variety of sensors such as pressure and touch systems, infrared rangefinders and accelerometers to determine their surroundings and location in relation to each other.

The LANdroids program currently is using iRobot’s Create educational and research robot for platform development and research tests. The prototype robots are powered by a Gumstick microprocessor and carry a commercial 802.11 radio package. He claims that the first-generation rugged LANdroid platform is on track and will be ready for operational testing in subsequent program phases.

Web Resources
DARPA: www.darpa.mil
Lockheed Martin: www.lockheedmartin.com
iRobot: www.irobot.com