Articles Written by John Kwasnoski
John Kwasnoski has written several articles for the Green Light News, the newsletter for Michigan's Traffic Safety Resource Prosecutor (TSRP) at the Prosecuting Attorneys Association of Michigan. Below are those articles, divided between law enforcement and prosecutor topics.
Click on the heading to read the full article.
(Thanks to Ken Stecker, Michigan's TSRP for permission to post these.)
In many cases the prosecutor is faced by the challenge of a professional expert being retained by the defense, and the potential for a highly credentialed expert to be pitted against the police reconstruction witness. Prosecutors must clearly evaluate the strengths of their own witness regarding a true comparison to the professional witness regarding the relevant qualifications. Many police (sheriffs, state police) reconstructionists have more formal training, more field experience with active crash scenes, and a more current competence in the technology of reconstruction than some professional experts and engineers. In this regard the prosecutor should evaluate the potential attacks on the professional witness, which might include the fact that the defense expert:
- did not personally observe evidence at the scene.
- relied on police measurements/photos.
- did not speak with police investigator or civilian witnesses.
- did not visit the scene until long after the crash - evidence was gone; perhaps never went to the scene.
- does not specialize in MV crash reconstruction – generalist.
- has academic publications/experience not related to reconstruction of MV crashes - look over the resume carefully - your police reconstructionist is often better qualified by his/her experience.
- prepared/amended report in reaction to prosecution’s report.
- may have a bias based on the fee being paid by the defense.
- is former police officer who has no better credentials than state’s witness.
- did not have anyone check his/her work for potential mistakes or errors.
- must confirm that reconstruction from evidence is more reliable than witness observations in most cases.
- used assumptions for calculations; the results are only as valid as the assumptions: were the assumptions chosen from the extreme end of a range to favor the defendant ?
- could not reproduce conditions at time of crash for later testing.
- will confirm investigative activities were correct (affirmative cross).
- will confirm that police calculations are mathematically correct.
- did not reconstruct the crash, but only finds fault with police reconstruction.
- may have to agree with a hypothetical that confirms the negligence of the defendant.
- can be impeached by prior testimony (do your research, and get those transcripts).
- has never worked for prosecution.
- has a relationship with this defense attorney, and is therefore biased.
The professional resume in many cases is not a match for the prosecution’s police witness who reconstructs MV collisions on a regular basis, is specifically trained to do so, and is a specialist within his/her agency. In addition, the issue of bias usually favors the state’s witness. In that regard it is important for the prosecutor to establish the credibility of the state’s witness during the direct examination by highlighting any investigative activity that would show a fair and unbiased investigation, including; seeking potentially exculpatory evidence, collecting evidence according to established protocols, and making multiple measurements to ensure fairness and completeness, etc.
The author has observed in many cases that the professional expert is no match for the state’s witness because jurors can make the most crucial decision in their comparison of the experts - who is more credible.
Many defense crash reconstruction experts utilize very “glitzy” computer programs to do calculations and generate convincing graphics. While the State’s reconstruction calculations, often done by hand, yield a true and accurate analysis of the defendant’s vehicle speed, direction of travel, availability of evasive action, etc. the defense may have a superior “dog and pony show” in terms of case presentation to the jury. Whenever possible, prosecutors should seek an admissibility ruling from the court in advance of trial.
A basic consideration for the prosecutor is whether the computer-assisted evidence will be offered as demonstrative evidence or substantive evidence at trial . The admissibility of the results of calculation or CAD software should be relatively easy to deal with. On the other hand, the more sophisticated software programs for reconstruction, simulation, and animation may be more readily challenged. The use of the computer-generated animation as a chalkboard that an expert uses to demonstrate his/her reconstruction testimony may relieve much of the burden of admissibility.
Some of the more frequently cited admissibility considerations include:
- Failure to warn opposition of intent to use animation and to disclose the technical contents of the software, including program listing.
- Evidence is redundant or repetitive.
- Computer-assisted evidence is prejudicial rather than probative (by its form it can have an effect which goes beyond its ostensible use).
- Calculations are beyond the capability of the expert.
- The computer results cannot be verified by the user as being accurate (in effect, the user has employed the software as a "black box" into which input evidence or assumptions have been blindly fed).
- Input information may not be in evidence.
- Software documentation (list of how the program calculates) was not made available to opposition prior to trial.
- The animation should be based on the results of reconstruction calculations and should not be "created" by the animator by placing vehicles at points or at speeds determined by the animator alone.
- Lack of input information requires "default" values selected within the software which do not match evidence in the crash.
- Judges may not understand the technology and therefore misinterpret the computer-generated results.
- Use of computer-generated results does not satisfy the Frye or Daubert criteria.
- Lack of appropriate foundation - user made assumptions that are interpreted by the software as physical evidence upon which reconstructions or simulations are based.
- Software was misapplied to a scenario for which the software was not designed (many reconstruction and simulation software program require a plane surface and will not accurately handle differences in elevation such as ditches, off-road movements, grades or super-elevations).
- Mischaracterizing an animation as "the way it was" rather than as a depiction of the expert's opinion.
- Failure to present testimony regarding the "real time" nature of an animation or the fact that the animation is done to a true scale.
The prosecution used a video animation in the case of TN v. Farner and obtained a conviction, but the Court of Criminal Appeals of TN at Knoxville (decided June 30, 2000) declared that the trial court had erred in admitting the animation at trial because the depiction of the motion of a third vehicle in the animation could not be supported by evidence and vacated the conviction. The Court of Appeals also said that the animation might be admissible if it could be corrected to fairly and accurately illustrate and explain the testimony of the officer. In this case the motion of a third vehicle was depicted without reconstruction of its motion from evidence at the scene.
 Kwasnoski, Partridge, Stephen, Investigation and Prosecution of DWI and Vehicular Homicide, Lexis Law Publishing, 1998
A multiple event collision has resulted in a fatality to the passenger; after skidding across the paved road surface, the car slid through the grass and struck a tree. The reconstructionist uses the conservation of energy principle to determine an equivalent speed for each event by using the speed from skid marks equation for the paved and grass surfaces, and then a crush analysis for the impact with the tree. A speed of 86 mph, well in excess of the posted speed is determined, and months later the case is in trial. All the prosecutor needs from the reconstruction witness is the speed testimony; the jury has seen the extensive damage to the vehicle, and knows how far the car was out of control prior to striking the tree. The jury sees the picture, but now the prosecutor makes a mistake that is not uncommon - he presents the speed testimony with poster boards full of mathematics, lengthy testimony about the calculations, and several hours of direct examination detail about the precision of the measurements and the gymnastics of the calculations that inject a techno-babble into the jury’s information gathering process. Not to mention the hours of cross examination by defense that is intended to confuse the jury, and to break down any connection the jury may have had with the expert witness. The defense then puts on their own expert who further tries to cloud the water by focusing on several of the details of the case that are nothing more than distractions. The result is four days of deliberation, and finally a verdict.
The moral of this story is really quite simple - LESS MATH IS THE BEST MATH. The jury is seldom, if ever, comprised of a panel of engineers, physicists, and mathematicians who can really understand the intrinsic beauty of the mathematical reasoning used to arrive at the opinion of speed. Rather, it consists of some people whose everyday lives do not intersect the world of mathematics at any deeper level than balancing a checkbook - which many find challenging. So why would the prosecution employ mathematical equations, and algebraic manipulations to convince the jury that the speed estimate is credible? The answer is simple - because that’s the way the witness talks, and many prosecutors buy into the rather impressive, although often mysterious, jargon and vocabulary of collision reconstruction. But convincing the jury, and being a credible witness should be the goals of the expert, not demonstrating the ability to calculate and spew circuitous definitions and theories of physics. Remember, a main reason a jury finds a witness credible is that they simply like the witness. The word nerd is not a term of endearment, so why not develop a strategy of making your expert likeable instead of simply competent.
In a recent case in Salt Lake City I testified about pre-impact tire mark evidence, crush damage that actually tore a vehicle in half, and post-impact motion that ended when the defendant’s vehicle jumped the curb and struck a house. The energy method of reconstructing the crash was to isolate each event, determine an equivalent speed to cause each event, and then to add the speeds together with the combined speeds equation. The mathematics involved several equations, and pages of mathematical details, but the testimony never mentioned a single number other than the final opinion of the speed of the defendant’s vehicle at the beginning of the events. In summary, the testimony of the opinion of speed consisted of an analogy:
The defendant’s vehicle had what we call kinetic energy - meaning energy because it was moving. And the amount of energy it had is directly related to its speed. So on the night of the crash the police documented evidence that the car’s energy had been lost during the collision. It was like knowing that I’d walked around and dropped coins on the ground - a nickel here, a dime there, and a quarter over there - and then someone asked the question how much change did John have before we got there. The investigator walks around the scene of my “coin tossing” and finds evidence of my activity, then adds together the observed coins, and opines, “John had at least 40 cents.” That’s what I did in this case - I converted the observations the police made into speeds that it took to do various things during the collision (skid, crash, rotate and roll, and strike the building), then I added the speeds together to find how much speed the car had to have to be able to do all the things. And when I did that analysis I determined the speed of the defendant’s vehicle had to be at least 74 mph at the start of this crash.
The testimony took less than half an hour, and jurors were smiling and nodding in agreement, obviously understanding the nature, if not the detail, of the method. During my direct examination I noticed another sign of success - the defense attorney turned over several pages of his yellow pad in what I interpreted to be anticipated areas of cross examination that he was abandoning because I hadn’t mentioned any of the detail of the calculation that he may have tried to explore. Since the cross examination is often a reactive process, there simply wasn’t much to which he could react. The direct was like a nice little story that confirmed what the jurors had already gathered in their observations of the damage photographs, the scene diagram presented by police, etc. The jury didn’t want a treatise on how to reconstruct the speed of the car, they just wanted to understand how it was done, and to have confidence in the accuracy of the speed. The defense did not call their own reconstruction expert, who had opined prior to trial that the State’s speed estimate was incorrect and flawed.
If the intent of offering expert testimony is clear, why cloud it by more detail than necessary. While it’s a hard decision for a prosecutor not to encourage the expert to display their ability, it may open the door for confusing cross examination or even worse - confusing the jury during direct. The judge acknowledges the expertise of the witness, so let the witness tell a story, not give a mathematics and physics lecture.
In a single-vehicle crash in a rural area a 2006 sedan leaves the road and strikes a tree, causing the death of the front seat passenger. The reconstruction report indicates an excessive speed, and the operator’s BAC indicates impairment; a charge of MV homicide is filed. There is no indication of any serious defense until several months later when the defense attorney enters a motion to have a defense mechanical expert inspect the vehicle for a possible failure/defect that could have caused the collision. While preparing a response to the motion the prosecutor learns that the defendant’s vehicle has been released to the insurance company, and has been destroyed. Can the judge dismiss the charge if the defense can show that a mechanical failure or defect defense would be feasible? How would the prosecutor explain to a grieving family that a prosecution cannot go forward because of a ruling by the court that potentially exculpatory evidence had not been secured.
The author has consulted on many cases involving mechanical failure and operator ID defenses in which the release of a vehicle seriously hampered the ability to respond to a defense theory of the case; and in two of those cases the charge of MV homicide was dismissed. For many agencies vehicle storage can be a serious problem because of cost or space considerations, but a vehicle is no less important than the weapon in any other homicide case. One agency started using a self-storage facility that is rented on a monthly basis, and secured by padlock, at an affordable cost. Another set aside some space at a Public Works lot and secured it with a chain link fence; the savings in the first year paid all the costs for fencing in the area.
The prosecutor and law enforcement should develop strategies for securing information from a vehicle, including:
- Processing at the scene – photographing the vehicle before it is moved, noting tire condition, damage, and inflation, and sufficient documentation of damage to prove that none has been done during transport. In some cases the lack of damage may be just as important as the damage itself, so the vehicle should be photographed all around its perimeter.
- Covering the vehicle for storage – this can simply be the use of a tie-down tarp to protect the vehicle from weather and add some extra degree of security.
- Doing a mechanical inspection – a routine inspection in which a qualified person inspects tires, brakes, steering, etc. is a good policy. And remember that the person doing the inspection may become the center of your case in a mechanical failure/defect defense when a highly-qualified defense expert enters the case.
- Checking for vehicle recalls – in every motor vehicle case every vehicle should be researched for recalls that may have been causative. Recall info can be found at the www.nhtsa.dot.gov web site. Officers might be encouraged to include such research in their report of the crash since it not only shows completeness, but also indicates a non-biased effort to find potentially exculpatory evidence.
- Photographing the vehicle interior – this may become critical in a subsequent operator identification defense. With digital cameras so readily available there should never be a shortage of photographs
- Downloading the CDR  – every vehicle that has a CDR should be downloaded, if possible, with a warrant obtained, if necessary. The failure to download a CDR can be portrayed to the judge and jury as an effort to hide potentially exculpatory evidence. When the CDR cannot be downloaded the reason may be included in the collision report to show the completeness of the investigation.
- Having a policy for releasing the vehicle – law enforcement is well-served by developing a policy in concert with all agencies for releasing a vehicle only after the prosecutor knows that the defense will have no further need to inspect the vehicle. Such policy must be respectful of individual agencies’ policies and politics, but releasing a vehicle prematurely is an open invitation to a mechanical failure/defect defense, and it can portray a lack of completeness in the investigation of the crash.
Once the issue of releasing vehicles has been considered it may take a lengthy discussion for the prosecutor and law enforcement to develop a joint policy that works for everyone involved. But in a motor vehicle homicide case the vehicle is a lethal weapon, and it should be secured just like the weapon in a murder case.
 The Crash Data Retrieval System (CDR) is an Event Data Recorder manufactured by the Vetronix Corporation. New and evolving technology is the norm in today’s society, and traffic crash investigation is no exception. The “Black Box” or Event Data Recorder (EDR) is a recently available technology that enables the investigator to access accurate and reliable data from a vehicle involved in a traffic crash.
Perhaps the most troubling defense in a motor vehicle homicide case can be one involving operator identification. A prosecutor calls or e-mails me with a panic tone because the superb collision reconstruction showing the alleged operator to be traveling at an excessive speed, or crossing the center line, or yawing off the road into a tree causing death to a passenger is challenged with regard to who was operating. Proving who was operating was not a part of the investigation, because investigators at the scene had concluded that they “knew” who was operating. Knowing something and proving it are two different things. Some of the investigative activities that make the operator ID defense less effective might include:
- anticipating the operator ID defense in every case, immediately, at the scene
- process the scene was processed with regard for proof of operation
- produce a scale drawing
- do a reconstruction showing vehicle motion to final rest
- do a vehicle inspection, including interior of the vehicle
- photograph the vehicle interior
- do a forensic examination of the vehicle, if there is any indication of operator identify problems
- do an autopsy
- collect samples of blood, hair, etc. from the alleged passenger
- identify, and question, all of the witnesses who could identify the defendant as the operator were not identified, or were not questioned with regard to who was operating
- obtain the 911 records as part of the investigation
- get the Ems records showing who had contact with the occupants of the vehicle
- save the clothing of the victim and defendant
- take photographs of the alleged operator’s injuries or lack of injuries
- secure and protect the vehicle
- do not release the vehicle without the prosecutor’s authorization
- check for potential evidence such as video cameras at gas stations, stores, etc.
- check for any evidence of purchases that might identify the operator
- include operator ID as part of every statement taken at the time of the crash
As just one example look at some of the forensic evidence associated with the vehicle that might be available to identify the operator, if investigators were aware of the potential need for proof:
- “pattern injury” on chest from steering wheel
- left side of head contact with A-pillar
- blood smears on interior of vehicle
- fingerprints on steering wheel
- fingerprint on key
- fingerprints on control levers, light switch
- fingerprints on rear-view mirror
- eye witnesses before or after crash
- blood spatter on driver’s side of vehicle
- knee injury from contact with dash
- seat belt marks consistent with belt use
- shoulder belt abrasions
- fabric fusion onto seat belt
- forensics on deployed air bag
- abrasion on forehead from contact with head liner
- forensics from windshield “spider web” fracture
- seat position
- damage to brake pedal consistent with leg injury
- pedal impression on bottom of shoe
- clothing fusion onto seat belt
- clothing fusion onto seat
- fabric fusion imprint on dash
- shoe transfer onto console (left-to-right ejection)
- inability of passenger to operate manual transmission
- clothing fibers in broken parts of dash, controls
- injuries to ribs consistent with striking door panel
- lacerations on face from windshield contact
- “dicing” from side glass implosion
- teeth impressions on vinyl dash material
- damage to rear-view mirror consistent with head injury
- “pattern injury “on leg from shift lever
- “pattern injury” on leg from door handle
- personal belongings under seat
- gas purchase receipts
- hair embedded in fractured windshield
- blood spatter evidence
Of course the investigator should not expect to find every item on the list or even many of them in any individual collision, but one or more could help corroborate other evidence of operation. Not finding any of them should certainly not be interpreted as proof of any kind.
This is a situation that prosecutors and law enforcement can work on ahead of time. Realizing that every case may be vulnerable to such a defense attack the prosecutor should meet with local law enforcement to establish policies to overcome some of the shortfalls in the standard investigation protocol; additionally, the prosecutor can ask that during training investigators be made aware of the critical nature of obtaining proof of operation in every case.
Over time the operator ID defense should become less and less problematic, but it won’t happen by itself.
Prosecutors and police have all seen the fatal DWI crash in which the defendant’s vehicle barreled out of control at a speed greatly in excess of the posted limit, went out of control, struck a tree or utility pole, and may have even torn the vehicle in two. There is little doubt about the speed, but as the case is prepared for trial the prosecution is troubled by the jury’s potential inability to find the defendant’s behavior to be reckless. So let’s take a look at what reckless operation of a motor vehicle really is from the perspective of the person who designs the roads to be safe.
A General Definition of Reckless
Webster’s definition of reckless includes such language as “not regarding consequences” and “irresponsible,” but making it clear to the jury might include relating the defendant’s behavior to the driver behavior for which the roadway itself was designed to be safe. Why was the speed limit of the road posted as it was - what safety and human factor considerations led to the decision to post the legal speed limit at 35 mph ? This could involve the town engineer or highway engineer, or an outside roadway design expert to explain to the jury the design considerations involved with the determination of a safe speed limit for any road. If the road is posted with a speed limit of 35 mph it should be understandable that operating at a speed of 65 mph on that road might create situations that are not safe, and that might endanger other people using the road.
For example, in the design of a new condominium complex the planners had to look at how much sight distance would be afforded to people in the complex who wanted to exit the driveway and enter the roadway safely. A sight assessment was conducted, and then a determination was made of the safe operating speed consistent with that sight distance. Perhaps changes were made to the road environment to provide the needed sight distance. If there was insufficient sight distance it may have been necessary to post signs on the road warning of a “hidden driveway.” The driveway design is evaluated with regard to established highway design guidelines published in either a state highway design manual or in a nationally-recognized manual like the “green book” . The professional highway design engineer can explain to a jury the consequences of people exiting a driveway onto a roadway when available sight distance does not allow them to see approaching traffic, because the traffic is traveling at too great a speed. The engineer can explain the reality of “an accident waiting to happen” when drivers operate at speeds well in excess of the posted limit at particular locations along the roadway on which the defendant operator traveled.
In one case in which the author worked, a site map of the roadway showed over thirty potentially dangerous situations created by the defendant operator’s excessive speed, including inability to see around turns in the roadway, over the crests of rolling hills, approaching pedestrian crosswalks, and approaching traffic control signs and intersections. Clearly, the design guidelines showed that at the speed the defendant was operating the situations were not safe for other drivers operating prudently. In fact, based on the defendant’s speed being so far in excess of the safe design speed for the road, the jury could clearly see that it was almost a certainty that the defendant driver would eventually cause a crash.
A site map could be used to show potentially dangerous situations where the defendant’s speed created a potential for disaster. The local engineer could tell the jury why each situation was so dangerous based on the guidelines used to design the road and determine what the speed limit should be. The jury should be able to see why they themselves would be in danger if they had been on that road at the time the defendant’s crash occurred. If it looks, walks, and sounds like recklessness be sure the jury can connect the defendant’s reckless actions with the legal definition they will hear in the charging instructions. This tactic of connecting the meaning of reckless to the safety considerations governing safe road design might resonate with jurors and give them a basis for reaching a decision.
[1.] “A Policy on Geometric Design of Highways and Streets,” 1990, AASHTO (American Association of State Highway Transportation Officials).
Law Enforcement Topics
In a post-impact movement in which a vehicle is rotating there is drag from the simultaneous rolling and lateral sliding movement of the vehicle as the tires experience the vector friction force created by the roadway surface’s reaction on the tires. This combined rolling and sliding effect is sometimes referred to in the context of the “friction circle,” which is a term of art referring to the vector component nature of the friction force on each tire of the vehicle. If a vehicle struck on its side goes into rotation and its orientation with respect to its direction of motion changes from 90o (moving directly sideways) to 0o (rolling directly forward) the drag factor on the vehicle would continuously change from full friction to rolling friction; there might also be an engine braking effect which would cause some drag. This effect and the resulting drag factor is reported in the literature (1,2,3,4) and is often used to determine a post-impact speed for each vehicle. In order to determine the correct drag factor to apply to the post-impact motion the varying drag factor must be analyzed as follows:
Determine the total rotation of the vehicle from POI to final rest position.
1. Break the post-impact motion into segments.
2. Determine the angular orientation of the vehicle with respect to its center-of-mass motion for each segment.
3. Determine the vector friction components for rolling and lateral sliding for each segment.
4. Combine the rolling and sliding components and average for entire post-impact path of the vehicle.
5. Apply the average drag factor so determined to the total length of the post-impact motion.
If the tires of a vehicle are not locked and “skidding” in the post-impact motion to the final rest position it may be completely incorrect to assume the full drag factor is acting on the tires. Instead, the tire orientation with respect to direction of motion should be determined to analyze the degree of rolling resistance and lateral friction that applies to each tire. Assuming full drag factor will often lead, erroneously, to a high speed estimate, and this speed may become part of a momentum calculation that results in a high estimate of approach speed for the vehicle, and for another vehicle as well.
The post-impact motion of vehicles must be investigated as thoroughly as other parts of the collision since the post-impact speed may be an important part of energy or momentum analyses.
Additional Articles on Drag Factor
1. Sgt. Thomas Shelton and Victor Craig, “Translational Deceleration from Vehicle Sideslip,” Accident Reconstruction Journal, Jan/Feb 1995.
2. Raymond M. Brach and Russell A. Smith, “Tire Forces and Simulation of Vehicle Trajectories,” Accident Reconstruction Journal, Nov/Dec 1991.
3. Duane R. Meyers, “Post Impact Deceleration,” Accident Reconstruction Journal, Nov/Dec 1994.
4. William H. Pultar, Jr., “A Model to Determine Deceleration of Rotating Vehicles,” Accident Reconstruction Journal, Nov/Dec 1990.
The momentum of a vehicle is defined by multiplying its weight by its speed, and then giving the directionality of the motion a mathematical description as well. For this reason the momentum equations look very complex and much more intimidating that most other reconstruction equations. The momentum analysis is based on a fact that is derived from Newton's Third Law of motion: the total momentum of all vehicles or objects before an impact is equal to the total momentum after the impact.
There are eight variables in the general momentum equation and six must be known to calculate the other two. Usually the two unknowns are the pre-impact speeds of both vehicles. The momentum analysis is independent of any damage or energy loss that occurs during the collision, and it is therefore a method of checking on energy calculations. If enough information has been collected at the scene to do both energy and momentum analyses, the results of the two speed calculations should be in agreement, although they rarely give exactly the same results.
It should be obvious that the numerous values needed to make the calculations require very complete processing of the scene, and an appreciation of how the uncertainties in each of the values might affect the calculated speeds. As with every other accident reconstruction methodology the results of a linear momentum analysis are only as accurate as the data used as input to the calculations - the success of the momentum calculation depends almost entirely on the level of investigation and how good the evidence is at the scene. When the calculation is completed the reconstructionist should go back over the calculation and do a sensitivity analysis to check if possible uncertainties in the data will produce significant changes in the calculations.
In a near head-on collision one vehicle drifts across the center line and strikes an oncoming vehicle in a violent collision that results in the deaths of two people. There is virtually no evidence of the pre-impact directions of either vehicle, but in an attempt to determine the speed of the vehicle that crossed the center line the reconstructionist assumes the oncoming vehicle to be traveling at the posted speed with a heading of 0o, and assigns an approach angle of 185o to the vehicle that crossed the center line.The momentum calculations yield a speed approximately 13 mph over the posted speed at that location, and the operator of the car that crossed the center line is charged with a MV homicide.
Issue: Is the momentum calculation in a head-on collision a reliable means for determining the speeds of the vehicles.
The application of the conservation of momentum theory to head-on or near head-on collisions requires very accurate approach angle data, since the calculations are extremely dependent upon the approach angles. In this particular case changing the approach angle of MV#1 by + 1o would change the calculated speed for MV#1 by as much as 80%.
Daily (1) best warns in his text that, “we must take great care in establishing our approach angle(s).” The numerical example Daily uses shows that for a change in the approach angle of 1o for one of the vehicles, the speed calculated for one of the vehicles changed by 23 mph. Be very careful when applying the momentum equations to near head-on collisions, as the calculation is very sensitive to the approach angles; and without corroboration be expecting a vigorous cross examination of the estimated speed.
1 Daily, Fundamentals of Traffic Accident Reconstruction, IPTM, 1988, p. 235
Two cars approach an intersection, neither vehicle brakes prior to collision, and the post-impact motions of the vehicles include rotation and vehicle rollout to the final rest positions. This is a common scenario, and the reconstruction of this crash might be done by a linear momentum analysis in which critical assumptions are made without understanding the significance of those assumptions. The collision may be treated as though the vehicles were idealized point masses, disregarding their centers of mass, the effect of rotation on effective drag factor, and especially the path of the vehicle from impact to final rest. In the Accident Reconstruction Journal “Letters to the Editor” column for Nov/Dec 1993, a writer points out some of the assumptions made in momentum calculations:
- CM at impact to CM at final rest position (FRP) is treated as a full braking skid, even though tire mark evidence of full braking is not observed. Effects of rotation on drag factor are not considered, and full braking drag factor is used. This was addressed by Pultar  in a mathematical model based on the friction circle concept.
- Separation angles used in the calculations are determined by POI to final rest direction, without regard for the actual post-impact path of the vehicles or vehicle rollout.
- Approach angles are determined by alignment of vehicle damage at full engagement.
Without correctly determining post-impact motion parameters a linear momentum calculation may become little more than an exercise in algebra and trigonometry. A limited at-scene investigation may make it impossible to complete a momentum analysis; this is better than making invalid assumptions in order to be able to complete a mathematical calculation. An example of a complex post-impact motion with significant rollout is the Dayton, OH intersection collision (in which a van rolled over a pedestrian crossing the street) that frequented the Internet recently. Using a POI to FRP direction for the separation angle of one of the vehicles would have been completely inappropriate, as there was significant rollout to the FRP.
To demonstrate the effect of an erroneous post-impact analysis fourteen “staged” collisions reported by Smith and Noga  were analyzed. The data on the staged collisions included scale drawings of the collisions and the post-impact tire marks, measured drag factors, final rest positions of the vehicles, etc. In each analysis the separation angle was determined by using the POI to FRP straight-line direction, and the entire post-impact motion was treated as a four-wheel skid.
The result of each analysis was compared to the known impact speeds, and a percentage difference was calculated. Two calculations resulted in negative impact speeds, which had no meaning. In two cases the impact speeds were underestimated. In the other cases overestimates of impact speeds ranged from 25% to 401%. Cutouts of the vehicle profiles were given to several reconstructionists, and they were asked to align the damages to determine approach angles. In another test the same reconstructionist was asked to align the vehicle damages more than once, and variations were noted. Aligning damage profiles generally produced an uncertainty within a range of +/-7 degrees. The variation of 7 degrees in the approach angle produced significant errors in calculated impact speeds in some cases, especially those in which vehicle weights differed significantly. The analysis of the Smith and Noga staged crashes offers a rare opportunity to work with well-measured and documented crashes, and affords a test of the certainty of linear momentum calculations when bad assumptions are made in lieu of evidence.
The analyses showed that the at-scene investigation of the post-impact trajectories, relates directly to the accuracy of the momentum analysis. When using the linear momentum method, ask the following questions:
- What evidence is there of braking action by each of the vehicles’ tires?
- Does the post-impact trajectory include significant rotation and/or rollout?
- What evidence supports the determination of approach and separation angles?
- Has a sensitivity analysis been completed?
 W. Pultar, “A Model to Determine Acceleration of Rotating Vehicles”, ARJ, Nov/Dec 1990
 R.A. Smith and J.T. Noga, “Examples of Staged Collisions in Accident Reconstruction,” Highway Collision Reconstruction, ASME, November 1980.
Wouldn’t it be great if you could take a few pictures of a scene, go back to the office, and produce a detailed scale drawing suitable for court and reconstruction analysis? In the best of cases that’s what the technology called photogrammetry might provide. The science of photogrammetry is based on extracting information from photographs using the mathematical law that the spatial location of a point is definable if it is represented in at least two images. Photographs taken at a collision scene, or photographs of a damaged vehicle, can be scanned into a computer in which software can then analyze the digitized images. With the increased availability and popularity of digital cameras many photogrammetry softwares are now available to directly analyze the recorded digital images from the camera. In a cutting-edge method used by the Utah Highway Patrol, a camera mounted on the belly of a small model helicopter captures the scene photos from an unobstructed overhead view. Once the photos are digitized, measurements can be made, and if more measurements are needed at a later time the images can be revisited to get additional measurements without going back to the collision scene.
Photogrammetry may be particularly useful in refuting the claims of a defense expert witness who interprets tire mark or other evidence differently than the police investigators who made scene measurements and observed the evidence first hand. It may also be vital to retrieving unmeasured information such as yaw radius, tire mark angle within the roadway, or other data that could provide necessary information that was not measured during the initial scene processing. Photogrammetric documentation of vehicle crush profiles can also prove invaluable when vehicles involved in collisions are released and destroyed prior to the realization that crush evidence may be needed.
With regard to scene processing, the use of photogrammetry can produce scale scene drawings that compete with Total Work Station or other on-scene measurement techniques for accuracy and detail, with the advantage of saving time on the scene. In most instances, however, the tradeoff is a longer time required in the analysis work back at the office that balances the total time required with Total Work Station and other methods of documenting a scene. It has also been suggested to the author that large scenes may not lend themselves particularly well to photogrammetric documentation.
MacInnis and Siegmund authored a paper about the historical development of photogrammetry , and noted some advantages of photogrammetry, including:
- Shorter closure time for the scene.
- Volatile scene data may be preserved, but might otherwise be destroyed by diverted traffic, rescue vehicles, or salvage activity.
- Greater details of vehicle damage may be revealed.
- Scene details (including debris) may be so numerous that it becomes impractical to locate each one by field measurements.
- Photographs can be taken by a lone officer.
- Any objects visible in the scene photos can be measured if needed for subsequent litigation.
At the same time there are a number of reasons why photogrammetry may not be widely adopted for crash scene documentation, including:
- Cost. While the software itself is less than $1,000, additional costs include training, computers, digital camera, plotter, etc.
- Specialized training required. It is the author’s experience that only certain investigators adapt well to the computer skills required to use photogrammetry softwares, so that some people receiving training may not eventually go on to use the technology comfortably.
- There may be no need for more than simple measurements in many investigations.
- The need for enhancement of evidence locations (especially at night) prior to photography may lengthen the time needed before the scene can be opened to traffic.
- It may be easier to convince a jury about the reliability of a direct measurement than to educate them about the complex computer-based technique of photogrammetry.
Feedback from agencies using photogrammetry can be seen in an article in the Accident Investigation Quarterly  in which interviews with over a dozen agencies are reported. Many of the agencies either use or have tested the Photomodeler software produced by EOS Systems, Inc (photomodeler.com)
 MacInnis and Siegmund, “Applications of Photogrammetry to Accident Reconstruction.” Proc. Of the Canadian Multidisciplinary Road Safety Conference VI, June, 1989
 Cooner and Balke, “Use of Photogrammetry for Investigation of Traffic Incident Scenes, AIQ, Fall, 2002