SITIS Topic Details

Proposals Accepted:  
Program:  SBIR
Topic Number:  A10-175 (Army)
Title:  Robot Localization & Navigation for Night Operations in GPS Denied Areas
Research & Technical Areas:  Sensors, Electronics
Acquisition Program:  PEO Ground Combat Systems
  Objective:  The contractor will develop a multi-stereo thermal-imager-based sensor head that would provide outputs of both a 360 degree infrared (IR) image of the surroundings of the sensor head and would have sufficient integrated processing capability to create, store and output a 3-D map of sensor head surroundings recorded during movement through a dark GPS denied environment.The recorded 3-D map would be sufficient for autonomous localization & navigation of follower robotic platforms equipped with identical multi-stereo thermal imager-based sensor heads or for autonomous return navigation of the robotic platform equipped with the map-generating sensor head.
  Description:  The Army has a need for a multi-stereo thermal imager-based sensor head that would provide precise localization and navigation maps for robotic platform operations in dark GPS denied environments. Using such a multi-stereo thermal imager-based sensor head, a robotic platform could be teleoperated in a dark GPS denied environment. While providing the thermal images for robotic platform teleoperation, the sensor head would also accumulate, process and register stereo thermal images and create maps of the sensor head surroundings for future autonomous robotic platform localization and navigation operations. The 3-D maps could be transmitted back to and used by follower robotic platforms to autonomously retrace the path of the teleoperated robot or used by the teleoperated robot to return autonomously by the original teleoperated path.

  PHASE I: The contractor shall design, develop & demonstrate breadboard IR sensor-based hardware system that can record, analyze, compress & store sequences of images as the sensor head is manually moved within a laboratory environment. The contractor shall develop & document the full sensor system top-level design and describe IR images analytical processes that need to be fully developed for the IR-sensor-based localization & navigation system.

  PHASE II: Based on lessons learned from Phase I, the contractor shall fully design, develop, and test a thermal-imager, 360 degree multi-stereo based sensor head that performs localization & navigation in a dark, indoor, GPS denied environment. Such system will demonstrate the use of multiple sets of thermal imaging stereo cameras in a single sensor head to record, analyze, compress & store sequences of images as the sensor head is moved within a relevant indoor industrial warehouse environment.

  PHASE III: Commercial opportunities for this sensor head are for all-weather robotic material handling, night-time security tasks such as: automated perimeter surveillance, mobile intruder detection, driver assist technology, and early fault detection for industrial machinery. Phase III military application opportunities for this sensor head are in all-weather robotic material handling, MOUT surveillance, inspections, autonomous and manned ground extractions, and covert leader-follower convoys. Examples are autonomous field logistics, automated perimeter and facility security, Night MOUT Operations, driver assist technologies, and asynchronous navigation from shared routes.

  References:  
1. Robot Spatial Perception by Stereoscopic Vision and 3D Evidence Grids- Hans P. Moravec CMU-RI-TR-96-34 September 1996 http://www.frc.ri.cmu.edu/~hpm/project.archive/robot.papers/1996/9609.stereo.paper/SGabstract.html

2. Robot Evidence Grids- Martin C. Martin, Hans P. Moravec CMU-RI-TR-96-06 http://www.frc.ri.cmu.edu/~hpm/project.archive/robot.papers/1996/RobotEvidenceGrids.abs.html

3. A Bayesian Method for Certainty Grids- Hans P. Moravec, Dong Woo Cho working notes of AAAI 1989 Spring Symposium Series, Symposium on Mobile Robots, Stanford, Ca, 1989 http://www.frc.ri.cmu.edu/~hpm/project.archive/robot.papers/1989/890118.bayes.ltx

4. Low Cost Navigation in GPS Denied Environment Project- TARDEC-TTC Prime Contract No. W56HZV-08-C-0701, Work Directive 001 http://www.techcollaborative.org/default.aspx?id=TTC_Mar17_Release

5. Sensor Fusion for Intelligent Behavior on Small Unmanned Ground Vehicles- G.Kogut, G Ahuja, B. Sights, E.B.Pacis, H.R. Everett; Space and Naval Warfare Systems Center, San Diego SPIE07 http://www.spawar.navy.mil/robots/pubs/SPIE07_6561-70.pdf

Keywords:  Stereo Thermal Imaging, IR Imaging, Evidence Grids, GPS Denied Localization, GPS Denied Navigation, Robot, Driver Assist, UGV, Unmanned Systems
Questions and Answers:
Q: NOTE: Additional information from TPOC for SBIR topic A10-175 in response to FAQs from multiple inquiries received during SBIR Pre-Release period.

Clarifications:
1. Sensor interface & control messaging shall use open source standards & video output shall be NTSC.
2. The sensor vertical field of view shall be a minimum of 25 degrees. Sensor effective IR optical range shall be greater than 1000 feet.
3. Desired sensor size is driven by intended/potential platform applications, which might include both manned & unmanned ground vehicles (UGVs); smaller & lighter will be considered better. The phase II sensor shall be smaller than one cubic foot in volume & weigh less than 30 pounds.
4. The phase II sensor shall have sufficient memory & processing capabilities for real-time data gathering for map generation & video transmission for teleoperation control of robotic platforms traveling at speeds up to 50 mph. The map generation can be at non-real-time speed.
5. Due to the sensor intended use on battery powered robots, phase II sensor design shall focus on minimizing power requirements.
6. Minimizing the bandwidth required to transmit the sensor derived map back to the robot's operational control unit (OCU) shall be a phase II sensor design focus for the IR analytical processing. Objective is to require comparable bandwidth & provide comparable map accuracy as that derived & transmitted from a differential GPS system.
A: o
Q: 1. I don't have an IR-camera/output (in JPEG format) system. If I can receive your typical IR-output's in .jpeg by internet so I can do the image processing at home, then I believe I can develop the software of 2D or 3D identifying and tracking system in at least one year. Two years will be much more reliable. I need to hire a post-doctor and I am sure he or she will not just come for one year.

2. How long and how much will your Phase I grant be?
A: 1. The phase I delivery requirements are stated in the solicitation: "The contractor shall design, develop & demonstrate breadboard IR sensor-based hardware system that can record, analyze, compress & store sequences of images as the sensor head is manually moved within a laboratory environment. The contractor shall develop & document the full sensor system top-level design and describe IR images analytical processes that need to be fully developed for the IR-sensor-based localization & navigation system."
No government hardware will be provided for the phase I effort.

2. The period of performance and maximum allowable cost are defined in the Army SBIR solicitation.

Q: What percentage effort is expected for the sensor development vs localization/mapping algorithm tasks?
A: The percentage of effort for the sensor development vs. localization/mapping algorithm tasks will vary for prospective bidders, based on prior research & core competencies. Whatever percentages are required, the important consideration is the requirement to deliver a working prototype sensor, which includes the localization & mapping capabilities.
Q: Does the Phase 1 effort require 3D image and stereo map generation in the breadboard demonstration?
A: Yes, the Phase 1 effort does require 3D image and stereo map generation in the breadboard demonstration. However, the 360 degree situational awareness requirement only applies to the phase II prototype.
Q: Will the Phase II prototype be expected to meet the stated size and weight requirements, or can it be a functional prototype system but not in the final form factor since it is intended for use in an indoor warehouse environment.
A: The Phase II prototype will be expected to meet the stated size and weight requirements. The robotic platform selected for sensor integration & evaluation in the indoor warehouse environment will be capable of transporting such a sized sensor payload. Do not assume that the warehouse topology will be smooth surface. Robotic test track may include rough terrain. If the developed phase II prototype sensor requires sensor stabilization for effective localization & navigation, then the sensor system should provide such stabilization.
Q: 1) Obviously this will be used in GPS-denied environments; however, does this also assume that no IMU data/odometry will be available. That is, will this sensor need to generate its own odometry estimate as part of the map-building?

2) There is always a tradeoff in stereo processing of the camera separation baseline versus maximum/minimum distance estimation. Is there a desired minimum and/or maximum distance at which obstacles should be detected in the images? If so, this will strictly put upper/lower bounds on the camera separation, thus will also put upper/lower bounds on the footprint of various sensor system configurations.
A: 1. Assume that no IMU data will be available. If your developed sensor needs IMU data for map-building, localization & navigation, then include an IMU.

2. Desired minimum and/or maximum distance at which obstacles should be detected in the images is speed dependent (i.e., the teleoperator should be able to stop the robot prior to hitting an obstacle.). Sensor effective IR optical range shall be greater than 1000 feet (i.e. sensor shall provide an IR image sufficient for a human operator to recognize an obstacle in the delivered image & stop a robotic platform going 50 mph prior to the robotic platform impacting the obstacle.). The effective sensor range is required for human perception & obstacle hit-avoidance response, not for the localization & navigation map building analytical processes.
As of midnight September 1, questions for solicitations SBIR 10.3 and STTR 10.B will no longer be accepted.
To read the solicitation for full proposal preparation and submission details click here.

Record: 30 of 367