GIS and Robotics: Robotic Platforms as Tools for Spatial Data Collection and Consumption
By Benton Yetman
October 05, 2006
Classified Ads:
 |
Beyond the effective reach of GPS, surveyors have traditionally used hand laser range finders with Tablet PCs to collect interior space data. This process is very effective for capturing room geometries, but can be very time consuming and especially difficult in non-traditional architectures. The repetitive nature of hand measuring can leave you feeling like a machine by the end of the day. The solution: Don’t be a robot, use one.
 |
The raw product of the robotic survey is a high resolution map that defines the interior space down to millimeter accuracy. These high resolution floor plans provide a bird’s eye view of the interior spaces in a building, including what is actually in them. These data are then processed and attributed by the survey team on tablet PCs in the field, capturing information about room use and occupancy. The final product is a high quality map of the interior space data of the facility that is ready to integrate into a GIS.
Patricia McSherry, chief GIO at Langley Air Force Base, was an early adopter of the survey method. "We are continually looking for new and innovative ways to gather data in order to help base personnel make better decisions through GIS,” said McSherry. "In this case, our search took us to the cutting edge of the robotics world."
McSherry put the technology through its paces recently during a pilot survey at Langley. During the survey, the robot was able to capture data for over 100,000 ft2 of office space per day, including cubicles and other non traditional features.
That may be fast compared to traditional surveys, but you have to look deeper at the real value of this interior space data to fully understand the benefits. "There is immense value in this data," said David Berez, a principal at Post Office Editorial. "With this method, you can get a real-time snapshot of all your facility interests in a fraction of the time. You can also use it to validate existing CAD/GIS data to look for holes or discrepancies."
 |
Robotic data can also be used to visualize interior space in 3D for applications that require volume or proximity analysis. This extrusion capability provides decision makers with the ability to make more efficient matches between available space and those who need the space.
So if the robotic platform is smart enough to map its environment, what happens when it can recognize where it is on the map? The term "Spatially Intelligent" defines a robotic entity which can determine its location by comparing the spatial data it has in memory to those which it sees in real-time through a combination of onboard sensors. Once localized, it understands the space and can dynamically navigate its surroundings. Adding specialized sensors enables the creation of focused missions for the autonomous agent, capable of patrolling the building while sending back sampling data about the environment.
Web services allow the robot to query a GIS based on location, and establish a two-way flow of information. The result: a spatially intelligent field agent that can collect, consume and relay information in real-time.
While the true potential for these systems unfolds, some GIS industry veterans are willing to wager a guess at what the future holds. Visionary geospatial icon Terry Martin of ESRI explains: "GIS aware robotic systems can form the basis for a virtual neural network of sensors around the world. Further, those robotic systems can be coordinated with GIS to act on information."
 |
Whether contributing to your GIS or accessing data from it, spatial robotics are a rapidly evolving tool with enormous value for the interior space world.
|
Your Comments Post a comment All comments provided in this section are those of the individual who has created the post. These are not the opinions of Directions Media, its editors, staff or owners unless otherwise noted. Directions Media retains the right to edit or delete any comments posted herein.
|
|
||||||
| The Sick laser range finder on the robot featured in the article contains systematic error (around 1cm) in it's measurements. Therefore article's claim of a "high resolution map that defines the interior space down to millimeter accuracy" is questionable - even if somehow, the odometry drift of the robot could be perfectly correctable. |
||||||
|
||||||
| The use of techniques such as Kalman or particle filtering can combine multiple sensor inputs to produce a state model with an accuracy greater than that of any individual sensor. |
||||||
|
||||||
| There are hundreds of ways to correct that. For starters, it has tons of measurements for every location using that radar, which can be statistically processed into a more accurate floorplan. Second that camera on top might very well be equipped with a very accurate distance meter. And the radar only used to get a general impression. I don't think that radar will be the only measuring device on the robot. We would, obviously, be most interested on how it works. |
||||||
|
||||||
| Albert, there is little reason that the resolution of the sensor is the limit of the final map. Multiple hits on a single point make this explicitly true. Techniques to extract features such as straight lines make it even more a reality. |
||||||
|
||||||
| http://blogs.zdnet.com/emergingtech/wp-trackback.php?p=373 Benton is gaining a lot of attention with this new solution. Great buzz. |
||||||
|
||||||
| After glancing over the demonstration which shows the robot collecting data: it wasn’t clear how exactly data gets transferred from a sketch like floor plan into a space model presented. It would seem like there is quite a bit work involved in getting from point of having a collected point raster locations into a usable plan. Speaking from experience, this technology should be a very good fit for instances when a sketch-like plan is required or overall quick verification of the existing plans for major discrepancies. One thing that keeps hunting me is if there is an accumulated error in data collection due to changing location of the robot in space. It is not clear what sort of benchmark point is used and if there is more than one. Dmitry Shkipin, Associate AIA Prime Edge Asbuilt Services 1684 Decoto Rd., Unit 157 Union City CA, 94587 (510) 499-4148 Tel. (510) 794-4178 Fax www.PrimeEdge.com |
||||||
|
||||||
| Nice article, Ben! Dmitry, there is no cumulative error because the robot is constantly comparing its current position, as noted by the laser, with its calculated position, as noted by wheel slippage gyroscopically corrected. To learn more about our mapping robots, including the MapperBot used by Ben, see www.mobilerobots.com. Jeanne Dietsch CEO MobileRobots Inc |
||||||
|
||||||
| This looks like an idea with potential. Working in the mining industry there is scope for this idea for completing cavity surveys in underground mining areas where due to safety reasons people cannot access. The robitic device would however have to be toughened up a lot and get larger in size to cope with wet and very uneven footings. |
||||||
|
||||||
| Jeanne, Thank you for the clarification. I will look forward to seeing more to come in this direction. It seems like a promising alternative to a more expensive solution known as 3D laser scanning. Dmitry Shkipin, Associate AIA Prime Edge Asbuilt Services 1684 Decoto Rd., Unit 157 Union City CA, 94587 (510) 499-4148 Tel. (510) 794-4178 Fax www.PrimeEdge.com |
||||||
|
||||||
| Hello, I'm a 2nd yr student hoping to do my dissertation in GIS and Robotics. I was wondering if you would be able to help me. Do you have any information on how the robot links to the pda that you could send to me? I noticed that both are mobile units. k0540837@kingston.ac.uk |
||||||
