Appendix N: Emergency Vehicle Operation
Emergency vehicles (EV), such as fire trucks and ambulances, using lights and siren while moving along a route of streets to respond to a request for emergency services can adversely affect traffic flow in the vicinity of the emergency vehicle. This adverse effect can be momentary or it may take a very long time to dissipate after the passage of the EV. In addition to the adverse effect the EV can have on the traffic, the traffic can have an adverse affect on the time it takes the EV to reach its destination. Traffic signal preemption has been shown to improve the EV response time and reduce traffic accidents, but it adversely affects traffic flow.
Traffic signal preemption temporarily alters normal signal time allocations in some manner. It gives preference to the movement of the preempting EV through intersections, with limited regard to the effect the preemption may have on other vehicles on the traffic facility. The character and significance of the effect depends on the geometric characteristics, the size and complexity of the traffic facility, the traffic loading and traffic mix (which is usually dependent on the time of day), the intersection approach on which preemption is being requested, the number and characteristics of the emergency vehicles requesting preemption, and the signal control methodology and the signal timing patterns in use.
An analysis was performed as part of Phase 1 of the ITS Public Safety Program’s Emergency Vehicle Network Delay (EVND) Study with the goal of illustrating and assessing the suitability of an approach for evaluating emergency vehicle preemption impacts on subsequent responders using micro-simulation. The background and characteristics of the experiment, the process of setting up and running the test cases, and the results of this assessment analysis were detailed in the report on the work completed in October 2001: Plan and Assessment for the Study of Impact on Delay, of the Use of “Lights and Siren”, and Signal Preemption by Emergency Vehicles Operating on a Network.
Based on the results of the assessment analysis, it was concluded that the use of the CORSIM micro-simulation with appropriate extensions provided an adequate starting point for the evaluation of emergency vehicle preemption impacts on subsequent responders. A significant finding of the Phase 1 assessment analysis was that the modeling of the special behaviors, which civilian and public safety vehicles exhibit when they interact, is essential to simulating real-life conditions and estimating impacts.4
It was determined that this functionality should be made public to promote discussion and research. This capability was designed for research purposes and can be quite tedious for large networks or large numbers of EV. At this time there is no intention of creating a user interface for this functionality or supporting it further. Only limited testing of the EV capability has been performed. Lack of real-world data to validate this capability against has been a problem. Therefore, this functionality should only be used with the understanding that it has not been extensively tested nor has it been validated.
Modifications to CORSIM
In the following paragraphs, some of the modifications that were made to CORSIM to meet the requirements for the EV modifications are presented. First, CORSIM supports several new inputs that are input via an external file with an *.evd extension. For example, for each EV, users will specify if it has signal preemption capability. Likewise, for each signal, users will need to specify if it supports preemption. Link inputs were expanded to include specification of shoulder and median availability. Several new driver behavior parameters were defined, including auditory and visual detection thresholds, reaction time factors, and motivational factors. Several new vehicle movements that result from a driver becoming aware of an EV, given the driver's current speed, distance from the EV, current lane, and motivational factors are supported. The probability of a driver choosing one of these possible movements is included. (NOTE: No modifications have been made to TRAFED at this time. Inputs for EV specifications are handled by a separate *.evd file created by the analyst.)
One of the basic elements to the operation of EV within the network is that when an EV enters the network, other vehicles will become aware of the EV’s presence. To facilitate this operation, CORSIM was modified to determine which vehicles are aware of the EV, based on the distance to the EV and the reaction thresholds of the individual drivers. As an EV traverses the network, more vehicles become aware of its presence. Once a vehicle becomes aware of an EV, the vehicle employs movement logic that will override its normal movement logic. This logic depends on many factors, such as: whether the driver is cooperative or not, whether the vehicle is moving or already stopped, whether the EV is traveling in the same or conflicting direction as the vehicle, whether there is a shoulder/median, whether the EV is to the left or right of the vehicle, whether the vehicle can make a lane change or is stopped at the intersection (and thus cannot make a lane change). The result of this logic is to choose one of the movements specified in the document Operator and Driver Special Behaviors from the Phase 1 study mentioned earlier.
Under normal operation, CORSIM vehicles randomly choose where to turn based on user input turning percentages. This can lead to some undesirable effects. For example, a vehicle may choose to make four left turns in succession, which would cause the vehicle to drive around the block. An EV responding to a call must obviously take the most direct route possible. CORSIM has the capability to force individual user-specified vehicles to follow user-specified paths through the network. EVs take advantage of this capability and follow specified paths through the network. The vehicle specification file was modified for EV processing as documented at the end of this appendix.
CORSIM was modified to have a new “fleet” of vehicles that are EVs. The EV fleet is designated in the *.evd file with a fleet number of four. Also, the number of vehicle types was increased to 16. These two inputs to CORSIM are undocumented features and only explained in this paragraph. The same vehicle type could be a member of different fleets. For example, a 6.1 m (20 ft) automobile could be of the Auto, Carpool, or EV fleet (to simulate a police car).
A vehicle of the EV fleet uses different movement logic than the other vehicle fleets. For example, an EV is able to use any lane to perform any turn movement. It is allowed to drive on the median or the shoulder if one exists on the current link. CORSIM was also modified to allow an EV to travel in opposing lanes, otherwise known as “contra-flow.” EVs are allowed to drive through a red signal and they use special logic when traversing the intersection. TRAFVU was modified to display the enhanced movements of both EV and other vehicles responding to the presence of EV.
Specific Implementation Notes
The definition of the input file that adjusts vehicle behavior based on the presence of an EV is contained at the end of this appendix. The file is the Emergency Vehicle Data (EVD) file (also referred to as the *.evd file). The data in the EVD file may eventually be made into optional record types that could be incorporated into the simulation input file. For the time being, they are implemented as a separate file that CORSIM reads in when it is present.
The EVD file enables the user to specify when a vehicle becomes aware of an EV and when an aware vehicle cooperates with an EV. These are determined by setting percentages of vehicles based on the distance from the EV to the vehicle in question. A random number is drawn and if the random number is greater than the percentage, the aware or cooperative flags are set for the vehicle. The speed that the EV is traveling can reduce the distance at which a vehicle becomes aware of the EV, as has been shown in many studies. The distance that a vehicle follows an EV can also be set by the user in the EVD file. Finally, the existence of preemption equipment at a signalized intersection and the range over which it can receive a preemption call is set in the EVD file.
The definition of the path following capability is contained in appendix M. This pre-existing CORSIM capability makes it possible for an EV to follow a specific path from origin to destination. For example, the EV can start at the firehouse and travel on specific links en-route to a specific exit node that acts as its destination. The vehicle file was modified to allow inputs for the speed adjustment factor that an EV may travel faster or slower than the free-flow speed. Care should be taken in adjusting this parameter, however, as many governing agencies have set rules for EV vehicles to travel an adjusted speed based on the posted speed and not the free-flow speed (e.g., 16.1 km/h (10 mi/h) over the posted speed). Another modification enables the specification of the range of the signal preemption emitter on the EV. A distance of zero indicates the EV is not equipped with a signal preemption emitter.
The changes to CORSIM that model emergency vehicles are quite extensive. Basically, the EV follows most car-following rules but has some overriding rules. The EV travels on the left lanes of the link because its “goal” lane is set to the left side of the link. The EV’s free-flow speed for the link is increased based on the input data for the vehicle. The EV searches for vehicles on surrounding links and flags them as aware and cooperative if they fall within the specified distance and have random number draws that indicate that the vehicle should cooperate. EVs will stop at a red signal, but then they can discharge when the vehicles at the other approaches are cooperating. An EV can also use contra flow to bypass a queue of vehicles. The contra flow lane was simulated using a non-existent lane eight, which is to the left of all other lanes. The EV checks for oncoming vehicles and vehicles check for oncoming EVs, and both react accordingly.
A cooperative vehicle will have its goal lane set to the right lane of the link and have its free-flow speed set to zero. The effect is that vehicles move to the right and slow down. However, if a vehicle is already on the far left lane in a left-turn bay, then it will remain on the left side of the link and not move across all lanes of traffic. If a right shoulder exists and is free of other vehicles, the vehicle may use it. The shoulder was simulated by using a lane zero, which is to the right of lane one. Lane zero is the lane used by buses to simulate a protected bus station. A vehicle stopped at a signal is permitted to discharge from the signal even if the signal is red, if the EV is in the same lane as the vehicle and close behind the vehicle.
CORSIM inputs to specify if a median curb or shoulder exists are specified on the Record Type (RT) 80 of the CORSIM input file (*.trf file). By default, shoulders and median curbs do not exist on any links in the network. By setting column 56 to 1 on RT 80 of a specific link, shoulders will be available for use by vehicles moving out of the path of an EV. By setting column 60 to 1 on RT 80 for a specific link, a median curb will exist and EVs will not be allowed to cross the median curb to simulate contra-flow as shown in Figure 117. If the median curb does not exist (default condition), contra-flow will be used when appropriate. These two inputs to the TRF file are undocumented features and only explained in this paragraph.
Figure 117. Illustration. Emergency vehicle contra-flow.
A fixed-time signal can be preempted from its normal cycle if an EV with an emitter is within range and the signal has a receiver. The signal will allow a pedestrian phase to terminate and then the signal will change so that the EV’s approach receives a green signal. This will last as long as the EV is preempting the signal. When the EV exits the intersection, the signal will transition back to its normal cycle. If the signal is coordinated with other signals (i.e., an offset has been set), an attempt is made to return the signal to coordination. This operation currently only applies to fixed-time controllers.
Emergency Vehicle Delay File Format
The EVD file is associated with a specific simulation input file via the filename. If the simulation input file is named “casename.trf,” then the EVD file must be named “casename.evd.” The EVD file contains five sets of values used by CORSIM to maneuver an EV through the network: awareness, cooperation, speed, trailing, and preemptable signal values. Each set has two lines of text that can be changed to make the simulation behave differently. A line that starts with “/” will be treated as a comment and will be ignored. Section tags are required, and data without a tag will be ignored. Just the first letter of the tag is actually used: “A” or “a” for Awareness values; “C” or “c” for Cooperation values; “S” or “s” for Speed values; “T” or “t” for Trailing values; and “P” or “p” for Preemptable signal values. An example EVD file is as follows:
/ The first set of values governs the awareness of a passenger vehicle to an EV in the vicinity.
/ It must start with the letter “A” or “a” for Awareness.
/ The first line has the distances where a passenger vehicle becomes aware of the EV, in feet.
/ The second line is the percentage of vehicles that become aware at those distances.
Awareness
50 100 150 200 300 400 500 700 1000 1500
99 90 70 50 40 30 20 10 5 2
/ The second set of values governs the cooperation of a passenger vehicle with the EV.
/ It must start with the letter “C” or “c” for Cooperation.
/ The first line is the distance where the passenger vehicle begins to cooperate with the EV, in feet.
/ The second line is the percentage of vehicles that cooperate at those distances.
Cooperation
50 100 150 200 300 400 500 700 1000 1500
90 80 70 60 50 50 50 50 30 20
/ The next set of values governs the reduction in awareness due to speed.
/ It must start with the letter “S” or “s” for Speed.
/ The first line is the speed the EV is traveling, in feet/sec.
/ The second line is the reduction in the percentage of vehicles that become aware at those speeds.
Speed
10 20 30 40 50 60 70 80 90 100
3 6 10 15 20 25 30 35 40 45
/ The next set of values is the distance behind the EV that the passenger vehicle will travel.
/ It must start with the letter “T” or “t” for Trailing.
/ The first line is the distance the passenger vehicle will trail behind the EV, in feet.
/ The second line is the percentage of passenger vehicles that will maintain at that distance.
/ The next set of values is the distance behind the EV that the passenger vehicle will travel.
/ It must start with the letter “T” or “t” for Trailing.
/ The first line is the distance the passenger vehicle will trail behind the EV, in feet.
/ The second line is the percentage of passenger vehicles that will maintain at that distance.
/ The next set of values governs the signals that are equipped with preemption devices.
/ It must start with the letter “P” or “p” for Preemptable signal.
/ The first entry on each line is the node number, and the second number is the range of the device in feet.
Preemptable Signal
220 1500
Path following Vehicle Modifications
Modifications to the vehicle file (*.veh) were made for the EV project. Normal (non-emergency) vehicles are separated by "N" or “n” for Normal. If no delimiter is present in the *.veh file, then CORSIM assumes the vehicles are normal vehicles. EVs are separated by "E" or “e” for Emergency. Vehicle types may switch back and forth using the proper separators. An example of a modified VEH file is as follows:
/ The format is:
/ entry time (sec), entry node, path ID, driver type, fleet #, vehicle type #
/ EV additional codes: speed adjustment (ft/sec from normal), and Preempt emitter range (use zero if not equipped)
Emergency Vehicles
45 8200 2 10 4 15 5 1500
90 8200 2 10 4 16 5 1500
145 8200 2 10 4 15 5 1500
Normal Vehicles
190 8200 2 10 0 2
245 8200 2 10 0 2
Emergency Vehicles
290 8200 2 10 4 16 5 1500
345 8200 2 10 4 15 5 1500
390 8200 2 10 4 16 5 1500
4 It was determined that this functionality should be made public to promote discussion and research. This capability was designed for research purposes and can be quite tedious for large networks or large numbers of EV. At this time there is no intention of creating a user interface for this functionality or supporting it further. Only limited testing of the EV capability has been performed. Lack of real-world data to validate this capability against has been a problem. Therefore, this functionality should only be used with the understanding that it has not been extensively tested nor has it been validated.
