A case study on vehicle safety testing in indoor environments
In order to address the growing need to perform vehicle testing in enclosed environments where a satellite-aided inertial navigation system can’t function, OxTS has developed an Indoor Positioning System to work with the ground-based signal location system developed by Australian firm Locata.
Locata’s alternative to GNSS/GPS was conceived to overcome the limitations of a satellite signal-based navigation system. Key to Locata’s system is a time-synchronization capability, called TimeLoc, which allows its ground-based transmitters, known as LocataLites (LLs), to synchronize with each other to picosecond precision.
A network of LLs forms a GPS-like constellation of transmitters that allows signal-based positioning within a serviced area. These networks can cover deep canyons, indoor facilities and other challenging environments where GPS struggles to operate, delivering centimetre-level accuracy with high reliability and guaranteed high repeatability. Locata-based commercial networks can operate as an alternative or an augmentation to GPS.
When an OxTS INS is configured to work with the Locata system, the built-in GNSS information is replaced by measurement input from the Locata receiver to produce accurate and reliable measurements while maintaining excellent position accuracy. Data is output via Ethernet and CAN to be used by other equipment, such as driving robots, or logged. Raw measurements are also logged internally to be downloaded for post-processing.
OxTS Indoor Positioning/Locata/AB Dynamics case study
One organisation facing up to the challenge of collecting accurate position measurement data under challenging conditions is the Insurance Institute for Highway Safety (IIHS). This independent American body conducts tests to assess how ADAS technology can prevent or lessen the severity of crashes. Several years ago IIHS identified a growing need to expand its test facilities while meeting the requirements for future testing, including all-weather operation and test automation.
The Locata network at IIHS’s Vehicle Research Centre uses 16 LocataLite transmitters. LL1 was designated as the master for the network. Shading denotes HDOP quality in the serviced area.
In 2015 IIHS completed a $30 million expansion of its Vehicle Research Center (VRC), the centrepiece of which was a five-acre covered track designed to allow testing to continue in all weathers. An existing outdoor track was also expanded, bringing the total test area to 15 acres. Given the need to simulate crashes safely, accurately and repeatably, IIHS has also researched the use of robotic equipment to automate some of the driving tasks.
At the VRC, IIHS operates a multi-frequency GNSS base station with real-time corrections to provide the position, velocity and time (PVT) parameters that are required for testing and essential for operating robotic test equipment. However, tests on the new covered track showed the equipment was not delivering the accuracy and repeatability needed, and it was concluded that the steel trusses of the covered track’s fabric roof were obstructing the GNSS signals. Finding an alternative technology that could deliver the required positioning performance on both the open and covered tracks triggered a global technology search, the result of which was a collaboration between OxTS, Locata and AB Dynamics, the world’s leading supplier of the driving robots used in automotive testing.
Test vehicle’s manual controls remained accessible to the driver despite AB Dynamics’ driving robot. OxTS RT1003 GNSS-INS and Locata receiver were mounted on the rear seat (not shown).
Driving robots precisely and accurately control a vehicle with a level of repeatability that vastly exceeds that of human test drivers. Historically, driving robots have been used for vehicle dynamics, durability and even crash testing, but when coupled with an accurate position measurement sensor they can execute centimetre-accurate path-following tasks. When developing ADAS the ability to accurately and precisely control vehicle position is key to recreating real-life scenarios.
AB Dynamics’ path-following software is an established and proven technology. Motion data is collected from an inertial measurement unit at 100 Hz and fed back to the robot’s path-following controller. This controller employs a speed-dependent look-ahead algorithm that not only maintains the vehicle heading but also allows centimetre-accurate path control.
A Locata network was deployed at IIHS with 16 LLs covering both the open and covered test tracks. The network was designed to meet two key requirements: firstly, accuracy of 10 cm or better at 95% confidence, and secondly, a very high degree of repeatability with a service availability (meeting the above accuracy requirement) in excess of 95% of the time. All Locata receivers were expected to time-synchronize with GPS time and use the same coordinate systems as GPS. The network infrastructure was built and is maintained by IIHS with support from Locata.
In October and November of 2017 IIHS, Locata, OxTS and AB Dynamics conducted a demonstration at the facility. An RT1003 GNSS-INS was used to receive PVT data from the LL receiver instead of the RT’s GNSS receiver, and the setup could be run interchangeably with either a GNSS or a Locata receiver. Both the RT1003 and the Locata receiver were mounted on one of the vehicle’s rear seats.
AB Dynamics provided a flexible driving robot drop-in kit that was installed without modifications to the vehicle and with the steering wheel, throttle and brakes remaining accessible to the driver. The Locata antenna was fixed to a roof rack-mounted ground plane, approximately aligned with the centreline of the vehicle. A second Locata antenna was connected to a second Locata receiver to be used for post-processing accuracy analysis of the fixed baseline between the two antennas.
Test procedure, analysis and results
The test vehicle was driven in various driving patterns on both test tracks. Double lane changes (DLCs), conducted on both tracks, resemble the driving pattern needed for testing most collision-avoidance and lane-change features, while an S-curve driving pattern was used to simulate IIHS’s headlight evaluations.
Data analysis from two full days of testing focused on the accuracy and repeatability of the automated test setup as a complete system first and then Locata alone.
The foundation for a highly repeatable control system with positioning accuracy is a highly reliable Locata network that delivers repeatable dilution of precision (DOP) and a number of ranging signals at any given track location. Repeatability of the numbers of LLs seen and the horizontal dilution of precision (HDOP) was investigated for this purpose.
During the five repeats of the DLCs conducted at 45 km/h on the covered track, the number of LLs seen remained constant at seven, as expected. For the 20 km/h lap scenario on the open track, the number of LLs varied between eight and nine,with the drop occurring at one end of the lap. Slight variations in the times of the drops were caused by the varying speed of the vehicle during the turns.
Analysis of the 48 DLC repetitions from the covered track, carried out at a range of speeds from 10 to 45 km/h, revealed a high level of repeatability. In straight segments the control system was able to repeat all the runs with less than 4 cm of mean deviation. A mean deviation of 5 cm was seen in the turns due to the range of speeds and the increasing lateral acceleration at higher speeds. The standard deviation also followed the same pattern, remaining below 3 cm during the straight-line segments and increasing up to 5 cm during the turns. A standard deviation of less than 2.5 cm was seen throughout all parts of the scenario, demonstrating that the Locata/OxTS/AB Dynamics automated control system maintained a run-to-run mean deviation of 5 cm or better and a standard deviation of 2 cm during straight-line driving.
With the addition of a covered test track, the IIHS needed an accurate and repeatable measurement system. Locata was able to install a local network of LocataLites to form a GPS-like constellation of transmitters to provide centimetre-level accuracy. Using a specially configured inertial navigation system from OxTS and an automated driving robot from AB Dynamics, highly accurate position data and highly accurate and repeatable testing was demonstrated.