GPS World - May 2008 - (Page 32)

AVIONICS & TRANSPORTATION | Vehicle Integration Å 10-FOOT CHANNEL trial survey points. changes the final prototype implementation. The number of alert levels was reduced from four (which included lowlevel Scan and Benign) to the two highest levels (Danger and Warning) because the subjects found the four-level scheme confusing. A central four-light Å FIGURE 5 Diagram of test course cluster was eliminated with survey points. because the subjects found it confusing when used with the LED string. The directional LED string was simplified to three light positions (Left, Center, Right) mounted on the dash directly in front of the driver with a width a little more than the steeringwheel diameter. These lights are used only for the Danger level. The Warning alert consists of audio only (the condition change triple-beep sound plus the verbal prompt ª Look for pedestrians nearbyº ). The Danger alert consists of verbal ª Pedestrian Right/Center/ Leftº played at a faster speed than it was originally spoken, and a single light to indicate the direction to the pedestrian. Radio Link Tests. The radio link integrated into the PAS must show adequate range and capacity to handle multiple peds. We performed tests in both an urban and residential setting to qualify the system in a realistic environment. For these tests, a test course with waypoints was identified using Google Earth. The urban course waypoints were about 100 feet apart and spanned up to approximately 1,000 feet. The residential environment waypoints were about 200 feet apart and spanned up to approximately 1,500 feet. (A quarter of a mile was the maximum desired range because longer ranges would activate P-units in no danger and occupy the V-units needlessly.) The range achieved in the residential tests (1,200 feet) allows 13 seconds of tracking before collision at 55 mph. This exceeds the 8-second system requirement. The range in the urban tests was shorter, as expected, but still allows 8 seconds of tracking at vehicle speeds of 40 mph, which is realistic for the urban canyon where the test was performed, as there are stop lights at each block. Co-channel transmission tests showed that for equal distances 32 GPS World | May 2008 each transmission has an approximately equal chance of reception and sometimes both messages are received, especially in crowded urban environments. The closer a P-unit is to the V-unit the more likely it will be received, which effectively implements in the radio a prioritization of more in-danger peds. Capacity tests showed the capability to adequately track a dozen P-units with 14 time slots, especially when considering that the closest units' transmissions are the Å FIGURE 6 PAS and kinematic most likely to be received. carrier DGPS positions relative We think we could get to the P-unit significantly more than 14 time slots in a production unit because we had to accept a very long preamble for each message (much longer than the message itself) that was fixed in the COTS radios we used. Also, lost data performance in the range tests and capacity tests could be significantly improved if we could implement forward error correcting (FEC) coding. The manufacturer of the transceivers used in the prototype had implemented only a simple parity check of the message bits. If this test was failed, the message was discarded, and we could not get access to raw bits of the message, so FEC was impossible. Differential GPS Testing The V-unit uses a custom designed, carrier smoothed CA-code Kalman ﬁlter. It produces its solutions from smoothed pseudoranges of satellites common to both the P-unit and V-unit. The test was conducted on remote parking lot at a shopping center just outside the D.C. beltway. A course as shown in FIGURE 5 was surveyed using a kinematic differential carrier-phase instrument called AutoDOCS-GPS that the authors developed for NHTSA to survey car crashes (ª Reporting from the Scene, Automated Crash Documentation,º GPS World, July 2003). It has a 1-sigma error of 2 centimeters. The stars in Figure 5 are arranged on each side of the center line to form a channel for the test car to pass through. A P-unit was placed on each side to test the ability to locate peds as they get closer to the car path. For safety reasons, the P-units www.gpsworld.com http://www.gpsworld.com

Table of Contents Feed for the Digital Edition of GPS World - May 2008

GPS World - May 2008 Contents Out in Front Letters to the Editor Expert Advice GPS III Contract Award Now a Reality? Near-Space Location Boost RTK Crops Up in Precision Ag Safety Afoot 50+ Leaders to Watch Making a Difference with GPS Product Showcase Advertisers Index & Company Directory Marketplace Classifieds Seen & Heard

GPS World - May 2008

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.