The need to cost-efficiently monitor heavy goods vehicle movements to and from the ports of Los Angeles and Long Beach motivated Caltrans to look at whether vehicle length could be assessed using single induction loops.
District 7 of Caltrans has been conducting trials to see whether the use of a single inductive loop to measure vehicle lengths and so identify heavy trucks is feasible. So far, the results have been very promising, according to Lead Transportation Engineer Steve Malkson.
Between them, the adjoining ports of Los Angeles and Long Beach, the US's two biggest, cover some 10,700 acres (43km2) and 68 miles (109km) of waterfront.Combined, they rank as the world's fifth biggest port: according to FY2008 figures, the Port of Los Angeles handled 170 million metric revenue tonnes (7.8 million 20ft equivalent units (TEUs); meanwhile the Port of Long Beach, according to FY2007 figures, handled 87 million metric revenue tonnes (7.3 million TEUs). That cargo has a combined value of US$380 billion and the vast majority of it, some 80 per cent, is moved from the port by road rather than rail. As a consequence, heavy goods traffic places a huge burden on local road infrastructure.
Like many departments of transportation across the US and elsewhere,
Identifying heavy vehicles
Traditionally, there are three basic methods of classifying and counting heavy vehicles. The first, Weigh In Motion (WIM), is prohibitively expensive when the need is to monitor large numbers of lanes and locations. The second, Automated Vehicle Classification (AVC) using piezo-based sensors, is less dear but can still be reckoned to cost about $6-7,000 per lane. The third option is to conduct manual counts over a defined period and apply various factors to arrive at an average daily traffic figure. However, recently in Caltrans District 7, which covers the city of Los Angeles, Ventura County and the two ports, tests have been carried out which have looked into the feasibility of using a single inductive loop per lane to accurately count the number of passing vehicles which are over 40 feet in length.
"We have to cope with huge numbers of bulk containers moving by road and have a real need to gain a better understanding of exactly how many there are, where they are headed and at what time of day," says Steve Malkson, Lead Transportation Engineer with District 7. "The issue is that we simply do not have enough money to pay for WIM or AVC systems in anything like sufficient numbers for the task at hand.
"We do, though, have upwards of 2,000 ramp metering installations. Our project looked to find out if, by getting existing controllers to work in tandem with an additional form of vehicle monitoring, it would be possible to gain sufficient information to a sufficient level of accuracy."
Caltrans issued a Call for Submissions at the
The two freeways in the Los Angeles area with the most heavy vehicle traffic were chosen to be the subject of studies. The first, I-710, is the main north-south route running from the ports to central Los Angeles. At the busiest point it carries almost 31,000 trucks per day. The second, State Route 60, runs east to west from central Los Angeles to inland southern California where it joins the I-10 and continues across the US. At its busiest point it carries around 28,000 trucks per day.
Two into one
Normally, two inductive loops are used to gain vehicle length data. The first loop gives an initial time signature. When the second loop is triggered, a simple time-versus-distance calculation gives the vehicle's speed. Measuring how long the first loop remained active against calculated speed gives vehicle length. With a single loop, the speed calculation is missing - all one gets is an indication of occupancy. Therefore, an assumption is made of average vehicle length in an entire mix of traffic. If a loop is occupied for longer than the average, it is assumed that a heavy goods vehicle has been detected.
To see whether the assumptions were correct, test sites were set up at three locations where AVC was already permanently installed in order to allow comparisons to be made. The results, says Malkson, were not perfect, but better than anyone expected.
"After some original set-up adjustments our original trials amassed two weeks'-worth of data working on a 24/7 basis. On most days, the system was accurate to within 5 per cent but there were days where that rose to 2-3 per cent. A prime driver of that accuracy is that the Wizard polls the loop at a much higher rate than do the controllers which we normally use. That means it can look at edge effects - effectively, it 'sees' the vehicle a lot sooner than a traditional controller.
Deployment potential
Next steps involve the installation of 26 additional systems at locations along I-710 and SR-60. This will be at locations where there are already two loops per lane and data from the original field trials will be used to calibrate these. Installations will take place in August and the server for the new systems was being installed as this edition of ITS International went to press.The final scale of deployment is very much dependent on the results from these expanded trials but if the systems work as hoped up to 200 locations in District 7 could be equipped. Malkson says that there are several thousand sites across Caltrans which could potentially benefit from the new system.
In addition to the technical innovation being displayed the PPP arrangement under which the whole exercise is being carried out is particularly notable, he says.
"As a part of the PPP, we provide InfoTek with a testbed," he continues. "That works well for both parties, in that Infotek gets a better product to sell and we get a better product to buy. Most of all, the whole thing isn't costing the taxpayers a lot of money but they do get a lot back in return."
The InfoTek Wizard
According to Malkson, installation of InfoTek's Wizard(tm) is especially straightforward. In a standard controller cabinet there is a harness with one wire per loop detector. A split connector goes to the 170 controller and to the Wizard. The only other addition is an antenna to allow the Wizard wireless communication.The InfoTek Wizard is a Java-programmable, GSM-based modem with 32 digital inputs and eight digital outputs. It offers bi-directional capability and an embedded Java runtime engine that executes custom applications locally. A number of features make it cost-effective in operation, the company says: its GSM/CDMA network capability reduces recurring costs; it offers the ability to stay connected but incur airtime charges only when data is transferred; and it provides a lower-cost alternative to dial-up and leased lines. Scalability allows hundreds of units to be deployed in days and it has demonstrated an up-time of 99.2 per cent in Caltrans installations, as measured by PeMS. The upgraded InfoTek Wizard II(tm) features: two serial RS-232 ports; an RJ-45 Ethernet port; dial-up capability; internal GPS; 24-channel digital inputs; and eight digital outputs.
For traffic data applications Infotek adds the Intelligent Loop Detector Application (ILDA), which polls the output of any loop detector card at 200 times per second. ILDA provides custom alerts in a variety of traffic conditions. It: features double or single loop configuration; eliminates manual downloading by polling automatically; captures volume, speed, loop occupancy and vehicle length; is compatible with most large database programmes; and supports up to 28 loops or 14 speed traps. ILDA makes use of InfoTek Data Collector(tm) software for traffic data collection, management and traffic alert notification. InfoTek Data Collector software can store the field data in most major databases, including MS Access, Oracle or MS SQL databases. The ILDA Report Generator provides customised reporting.
