SSC Pacific
Communication Relays: ADCR Gen-4



With the completion of MDCR, we returned our focus to ADCR, but this time with the goal of providing more autonomy for the fielded MDCR systems. With FY12 internal NISE funding, we demonstrated a new system that could deploy the MDCR relay nodes either automatically or semi-automatically. Motorized deployment forks were designed that were mounted on the rear of the robot (an iRobot 510 PackBot was used as the test platform—see Figure 6). A dongle was designed for the OCU to provide buttons for switching between automatic and semi-automatic deployment modes and for controlling the deployment of individual relay nodes. (The addition of the dongle ensured that the system remained plug-and-playable, with no modification to robot or OCU software.) The link-quality estimator that began under the third-generation ADCR project was completed. The estimator provides a “weak link” warning based on received signal strength indicator (RSSI) data and an “imminent failure” alert based on trends in the video data throughput. Through extensive testing, we noted that the video data throughput remained fairly constant long after that RSSI value had decreased. Then at the point where video glitches were beginning to appear, the averaged video data throughput dropped fairly linearly while its variance increased significantly. We designed two linear classifiers to detect this condition, one based on the throughput variance, the other on the averaged throughput trend. Supervised learning algorithms were used to train these classifiers. An “imminent failure” alert is issued when both classifiers detect the condition for three consecutive time samples.

The fourth-generation ADCR system on an iRobot 510 PackBot.  The left node is being deployed.
Figure 6. The fourth-generation ADCR system on an iRobot 510 PackBot. The left node is being deployed.

For this design, the OCU end-point radio performed the link monitoring function instead of the Deployer Module as in previous generations. This was necessary in order to monitor the throughput of the received UDP video data. In the automatic mode, the imminent-failure alert triggers the automatic deployment of a relay. In the semi-autonomous mode, a link-quality indicator LED on the dongle turns from green to yellow to give the weak-link warning and inform the operator to start looking for a good location to deploy a relay node. This LED turns red to indicate the imminent-failure condition, alerting the operator to deploy the relay immediately. We believe that this mode would be favored by the user in the field. It allows the user to control exactly where a relay node is placed. This location can depend on a number of factors unknown to the system, for example: the overall mission, upcoming maneuvers, the terrain ahead, visibility and ease of locating the relay node for later retrieval, etc.

For further information, see:

  • Pezeshkian, N., Neff, J.D., and A. Hart, "Link Quality Estimator for a Mobile Robot", 9th Int. Conf. on Informatics in Control, Automation and Robotics (ICINCO 2012), Rome, Italy, July 28-31, 2012. 
  • Hart, A., Pezeshkian, N., and H. Nguyen, "Mesh networking optimized for robotic teleoperation", SPIE Proc. 8387: Unmanned Systems Technology XIV, Baltimore, MD, April 25-27, 2012. 
Bookmark and Share
Updated: 7/15/2013 10:22 AM EST   Published (1.0)