|December 16, 2009 |
Remote Robotics and Innovative Mapping Techniques
The Design and Manufacturing Institute (DMI) at Stevens Institute of Technology has, since 2005, been conducting research on intelligent electro-mechanical systems under the direction of Prof. Kishore Pochiraju. Their efforts in this field have resulted in a Remotely Operated and Autonomous Mapping System (ROAMS) vehicle that maps an environment while returning a real-time, detailed 3D view of the location. ROAMS is an unmanned platform consisting of a sophisticated 3D mapping sensor, video cameras, microphones, proximity sensors mounted on top of a retrofit All Terrain Vehicle (ATV). Information is relayed via a patent pending wireless technology known as Sensor Manager for Adaptive Real-Time Transmission (SMART) to a custom built Operator Control Unit (OCU) Software system which provides monitoring, control and supervision.
ROAMS technology is unique in that it solves two of the fundamental problems associated with current mapping systems:
- ROAMS is dramatically cheaper than its competition.
- ROAMS platform is built to accomplish mapping missions with a great degree of autonomy.
Autonomy in the system is beneficial in exploring areas where little is known and the communication to a human operator cannot be guaranteed. First and foremost, its potential use in the military could greatly improve safety while generating situational awareness with mapping, environment sampling and relaying the information in real-time. ROAMS allows control over a wireless network, up to a mile away, to produce detailed 3D maps quickly, accurately and conveniently.
ROAMS maintains cost effectiveness by integrating several commercially available sensors and fusing the information generated by them in real-time. ROAMS makes use of a customized 2D Light Detecting and Ranging (LIDAR) system which is combined with cooperative technologies can produce 3D mapping scenes of equal quality and for far less money than commercially available 3D LIDAR systems.
The challenge for Dr. Pochiraju's team was to utilize low-cost 2D Lidar system that performs and design a 3D system that has resolution and accuracy to rival commercially available systems. To do this, they have mounted 2D Lidar onto a unique and adaptively controlled three-degree of freedom rotating platform. According to Mr. Biruk A. Gebre (BSME, 2006, MSME, 2009), Research Engineer at DMI, "by acquiring the simultaneous position feedback of the actuators during each 2D scan and using the actuators to change the orientation, 3D scans of an environment are created." This system is mounted on top of the ROAMS vehicle and provides a near 360⁰ unobstructed 3D scan.
Biruk must also integrate and synchronize many of the additional information gathering technologies onboard ROAMS. "In this system, a 2D Lidar, video camera, 3 servo motors and 3 angular position sensors are used to produce 3D color scans" – Biruk. The integration between these sensors and actuators allows the group to perform depth adaptive scanning and dynamic adjustments.
Digital 3D Scene Rendering
Hao Men is a Ph.D. student working with Dr. Pochiraju who specializes in making sense of the mapping data from the scans. Hao's work must deal with several challenges due to imprecise position and pose of the vehicle as well as non-unique solutions. In order to achieve better map accuracy and resolve ambiguities in map fusion, Hao's algorithm utilizes position from the 3D LIDAR as well as hue and texture information from the video streams. The map is actually a conglomerate of point clouds, individual coordinates mapped at a frequency high enough to portray a 3D image of the surrounding environment. Distances are represented by colored points on a sliding scale – the lowest vertical distances being red, while the highest are blue. The 3D scan makes use of the hardware installed by Biruk, and provides distance measurements in addition to reflectivity values, which help determine material components.
Autonomous Navigation and Vehicle Design
Working at DMI, Dr. Pochiraju and his team have retrofitted a miniature ATV as the frame for the ROAMS vehicle. This frame is constructed from high tensile strength hollow steel tubes and is connected to the all terrain wheels through the use of hinges and a front and rear suspension system. A secondary frame is mounted on top of the base frame to allow for easy mounting of hardware components. Mike Manzione (BEME, 2008, MSME 2010), a Master's student, studies the optimization of power management in the ROAMS vehicle. Large electrical battery banks have been installed on the vehicle, providing power from everything from motors and servos, to the on-board computers, sensors and wireless devices. ROAMS vehicle can operate continuously for more than 12 hours on a single charge.
In commercial and military applications, reliability is key. Wonyoung Kim is a Ph.D. student who specializes in machine diagnostics and health monitoring technologies. The team knows the importance of technology functioning successfully and efficiently, so Kim uses real-time sensing, modeling and simulation and knowledge bases to ensure the continued reliance on ROAMS vehicles.
Wireless Communication and Software Control
One challenge for Dr. Pochiraju's team has been the establishment of a better method of transferring data over a wireless network. Current wireless network technologies are limited by both latency (the time it takes for a signal to reach its destination), and throughput. The ROAMS vehicle is extremely complex and requires high bandwidth to function accurately. If latency between the ROAMS vehicle and operator becomes increasingly large, response times may suffer.
Their solution is a patent pending technology known as SMART. SMART technology will allow for dynamic adjustments to be made to the throughput of a system by enabling real-time adjustment of sensor data resolution (without drastic loss of sensor quality data).
ROAMS is an entirely unmanned remotely controlled method of surveying using an Operator Control Unit (OCU) software program, developed by Biruk, which provides a graphical user interface for mapping and positioning information and is used for remotely operating and monitoring the status of ROAMS.
Screenshot of the OCU software developed by Biruk A. Gebre
Taking ROAMS technology into the future
The establishment of an autonomous vehicle that records and presents advanced 3D mapping information has brought the work of Dr. Kishore Pochiraju and his team at DMI into the spotlight recently. Their work is garnering widespread attention from the media, and has been well received at conferences and showcases. Inventive methods of fabrication and control result in ROAMS costing a fraction of similarly equipped systems, and the autonomous nature allows for dramatically improved safety and analysis of unknown or hostile areas along with increase speed, accuracy and convenience.
Dr. Pochiraju's team is poised to go even further by improving upon their existing design. Chirag Gardharia, Master's student, is paving the way for new ROAMS innovations. His work involves development of a next generation, low cost, high scan frequency 3D laser scanner. Ph.D. candidate Akin Tatoglu also pursues this mission and researches control of linear and non-linear dynamical systems. In robotics, he reports, "optimal control and trajectory generation methodologies can benefit from mimicking biological systems and power and performance while efficiencies can be improved by miniaturizing the devices into micro- and nano-scales."
Mike Manzione is conducting power usage traces on the ROAMS vehicle, including real-time tracking and monitoring of the instantaneous power systems of the vehicle. He also specializes in the study and analysis of new, more efficient fuel sources. One area of particular interest is fuel cell technology, which he believes may be "a more efficient power solution based on peak load and desired run time criteria established by the power monitoring system."
Dr. Pochiraju's research has both offered solutions to key technology issues and served as an academic research platform for Stevens students. ROAMS technology is a conglomerate of disciplines and a testament to the ability to achieve dramatic advancements through coordinated research efforts. It combines patented technology with innovative and creative solutions to dramatically reduce the costs of survey mapping while enhancing its safety and efficiency and offers the potential to revolutionize the field.