On Feb. 14, the Google self-driving car caused its first crash. As the car moved into an adjoining traffic lane, its cameras spotted a municipal bus, but the car’s control software interpreted that the bus would slow, and that there would be no problem changing lanes. Unfortunately, the bus didn’t slow, and the car ended up sideswiping the bus.
In the Google example, the business intelligence built into the software can be amended to correct erroneous assumptions that the control software made about buses, but there are also deeper issues that impact disciplines like surveying. These issues focus around the effective integration of GPS, LiDAR, cameras and sensor technology.
First, let’s take a look at GPS.
Whether it is self-driving cars or surveying, maps on record don’t always reflect physical reality. Roads are modified, renamed, closed for construction, etc., and in some cases it becomes relevant if there are potholes or other hazards awaiting vehicles. In a survey project, these physical realities can play a major role if a surveyor is engaged to help chart a route for a truck trying to access a remote mining site over difficult-to-navigate terrain, or even in a case where surveying activities are taking place on urban streets or at a construction site with unforeseen hazards.
Carmakers propose to bridge this gap between maps and physical reality by capturing data on the fly through LiDAR and sensor systems on vehicles, and then sharing this data with other vehicles as input gathered from each car is amalgamated into a collective intelligence that any car can access. If a model like this works correctly, the first car through an intersection under construction will see that an intersection is under construction and that a lane change (and possible traffic congestion) results before a second car arrives at that intersection. This technically enables the second car to take corrective action to avoid the intersection altogether and to pursue an alternate route.
Cameras, LiDAR and sensors capture what actually exists and then enrich GPS/GIS information with the data. It is the ability to collect and transmit this data in near real time that encourages Google to advocate that on-car LiDAR, cameras and sensors detect potholes and other road obstructions.
Similar refinements to mapping and physical representations of sites can be made in surveying. Today, mobile LiDAR technology is being used to identify and repair potholes for the maintenance of public transportation infrastructure.
LiDAR software can also automatically classify ground features such as terrain shape\angle, terrain height (max\min), object height (min\max) and building height (max\min).
Surveyors are putting these technologies to work to cut time and to gain a more accurate portrayal of the sites they are working on.
In one case, Stanley Consultants, an engineering firm, performed design engineering for improvements of an intersection and roadway widening in an urban project. The area was multi-modal and well-traveled by pedestrians, bicyclists, transit systems and motorists. Stanley surveyors and designers used a combination of LiDAR and ground surveying to conduct a detailed right-of-way boundary and topographical survey. They did this by strategically placing a LiDAR station so it could optimally capture data, with surveyors using wireless remote controls to operate the system.
Projects like this exemplify the importance of advancing the integration of LiDAR, sensors, GPS and traditional on-the-ground surveying techniques to obtain a richer and more informational vision, whether it is of a route, a building site or open terrain. In this sense, land surveying and autonomous navigation systems on cars share the potential of technology integration that is still emerging. The Google car accident illustrates that this integration is still a work in progress, but as more is learned about LiDAR-sensor-GPS integration, the ability to gather, aggregate and analyze real-world and GIS/GPS information will also expand visualization in related disciplines like surveying.