R v Soni, 2014 ABQB 709

Court of Queen ’ s Bench of Alberta Citation: R v Soni, 2014 ABQB 709 Date: 20141119 Docket: 110247475Q2 Registry: Edmonton Between: Her Majesty the Queen Respondent / Crown - and - Jayant Soni Applicant / Accused _______________________________________________________ Reasons for Decision on Voir Dire #4 of the Honourable Madam Justice D.L. Shelley _______________________________________________________ I. Background [ 1 ] Mr. Soni stands trial for: 1. dangerous driving causing death ( Criminal Code , RSC 1985, c C-46 ss 249(4)), and 2. street racing causing death ( Criminal Code , s 249.4(4)).

[2] On April 23, 2010, a traffic collision occurred in Edmonton at 66 St and 31 Ave. which led to the death of Jeremie Lablanc. [3] The collision involved three vehicles: a 2005 Mercedes Cabriolet (the “Mercedes”), a 2007 Lexus IS250 (the “Lexus”) andan Oldsmobile Alero (the “Alero”). The Mercedes struck the middle of the passenger side of the Alero as the latter was turning left at the intersection of 66th Street and 31st Avenue. The Lexus veered to the right onto a grassed area off of the road surface. It sustained somedamage prior to doing so, although significantly less than the other vehicles. [4] Mr.

Soni was the driver of the Mercedes and Mr. Leblanc was the driver of the Alero. The Crown argues this collision andthe death of Mr. Lablanc resulted from Mr. Soni’s alleged illegal conduct. [5] On Nov. 4, 2010 an employee of Mercedes-Benz Canada downloaded information from the Mercedes’ onboard ElectronicControl Unit [“ECU”] system.

This information was provided to Edmonton Police Service [“EPS”] investigators and the Defence. [6] The Lexus also had onboard electronic data that was downloaded from that vehicle on Sept. 29, 2010 by a technicianemployed by Toyoto Canada, who then provided that information to the EPS. [7] In a voir dire, the Defence applied to have this evidence excluded. It argued that the admission of this evidence, in theabsence of a method to evaluate its reliability, was a breach of Mr. Soni’s rights under s 7 of the Canadian Charter of Rights andFreedoms,

Part 1, Constitution Act, 1982, being

Schedule B to the Canada Act 1982 (UK), 1982, c 11 [the “Charter”]. Accordingly, theDefence argued that it should therefore be excluded under the Charter, s 24 or pursuant to: R v Harrer, (SCC), [1995] 3SCR 562, 128 DLR (4th) 98. [8] On Oct. 31, 2014, I concluded in an oral decision that neither the electronic data from the Mercedes or Lexus should beadmitted at trial and considered by the finder of fact. I indicated at that time that detailed written reasons would follow. These are thosereasons. II. The Mercedes Data A.

The Relevant Data [9] The ECU data recovered from the Mercedes took the form of a printout of various data points recorded by the Mercedes’onboard electronic systems (Voir Dire Exhibit #4). The Crown sought to rely on one aspect of that data as relevant: that a single eventoccurred when the vehicle was travelling at 105 km/h.

That event occurred when the vehicle had a kilometer reading of 79050. [10] Rather surprisingly, the Crown did not enter evidence of the Mercedes’ odometer reading after the collision. [11] In brief, the Crown sought to have that data admitted as evidence of the speed of the Mercedes at the point of impact with Mr.Leblanc’s vehicle. B.

Evidence on the Recorded Data [12] Documentary and witness evidence relating to the ECU data, its collection, and nature was submitted in the voir dire. [13] Voir dire Exhibit #5 was an Oct. 14, 2010 letter of Anne Bell, Customer Relations Manager, to EPS investigator ConstableMartin Desmeules concerning electronic data systems on the Mercedes model involved in the collision. It explained that a number ofelectronic systems receive data from sensors. If a signal terminates or other circumstances arise that indicate malfunction then that datais recorded as a diagnostic code in a control unit. [14] Ms.

Bell continued to explain the intended purpose and limitations on this data: ... These stored diagnostic codes are used by workshop technicians to quickly localize the source of a disturbance and identify damagedcomponents. Providing diagnostic assistance was the only intended use of this data and is the only reason it is stored. Regarding the specific information available following an accident, as a general matter, an accident may or may not give rise todiagnostic codes that will be stored in vehicle control units. Whether codes are stored will depend on the circumstances of the individualaccident.

Whether codes are stored will depend on the circumstances of the individual accident. For example, if system components suchas electrical wiring are damaged as a result of the accident, a code indicating that that “fault” in the wiring has occurred may be storedbut these diagnostic entries are nothing more than snapshots of the latest changes to the system. However, the vehicle is not equippedwith an accident data recorder (EDR), which in case of an accident stores the accident-relevant data as for example the collision speed,the change in velocity or the deceleration values.

Diagnostic data stored in the Mercedes-Benz control units are generated at a particular point in time. A chronological assignment ofdiagnostic entries in different ECUs is not possible due to the fact that the ECU’s work with an internal and not a global synchronizedoperating time. In some control units, there may be some indication of the vehicle odometer reading at the time the event occurred which gave rise to thediagnostic entry. If such data is stored, it is only stored at relatively broad intervals. This means that it cannot be determined whether anentry was made before, during or after the accident.

In order to secure existing data that has survived the accident and ensuing rescue and removal operations, no power should be supplied tothe control units (e.g., by cycling the ignition). Rescue and vehicle recovery activities (e.g., towing) can change diagnostic data and/or, if

the ignition is turned on after the accident, new diagnostic codes may be recorded. Furthermore vehicle components may be damaged during rescue activities as wel l. [Emphasis added.] [ 15 ] The Crown called Antonio Barios, an assistant manager for technical services at Mercedes Benz Canada. Mr. Barios was the individual who in 2010 extracted and analyzed the ECU data recovered from the Mercedes. He appeared in the voir dire to interpret the ECU data but was not qualified as an expert. [ 16 ] Mr. Barios made several key points: 1.

ECU data is intended purely for diagnosing and repairing faults in a vehicle - it is not intended for accident reconstruction; 2. ECU data is recorded in relation to periods longer than seconds; 3. almost all ECU data is not ‘time and date’ stamped, and for this vehicle no time and date data had been recorded; 4. ECU information recording criteria are set by engineers, and those criteria are not available; 5. data recorded was either a “fault code” or an “event code”.

The former is a basis for later diagnosis of malfunction, the latter tracks communications interruptions between ECU’s; 6. the Mercedes Electronic Stability Program (ESP MK25) had lost a power supply on one occurrence, and that interruption had begun and ended at a point where the vehicle’s distance travelled reading was 79,050 km, and at that point the Electronic Stability Program had received data recorded by a different ECU that the Mercedes was travelling at a speed of 105 kilometers per hour; 7. he could not:

a) correlate the 79,050 kilometers travelled data with a time or a specific point in the vehicle’s operation, and

b) restrict when the power interruption event occurred in that kilometer’s travel; and 9. the ECU data may be altered by rescue activities or towing if the vehicle ignition is switched on, even if the vehicle is not running. [ 17 ] Mr. Barios candidly explained his limited ability to explain the operation of the ECU’s and

interpretation of the ECU data: 1. the Mercedes had more than one ECU, but he did not know how many; 2. he was not trained to interpret the ECU data; 3. he did not have access to the source code for the software used by the ECU’s, nor their operating instructions; 4. he could not explain how the vehicle’s sensor data was interpreted by the ECU’s software to obtain the recorded information; 5. he could not evaluate the accuracy of the data stored on the ECU’s, or the error rate in their operation; and 6. he could not determine whether the data recorded was at the time of the accident or some point before or after. [ 18 ] Mr.

Barios generally agreed that Ms. Bell’s letter was accurate. III. The Lexus Data A. The Relevant Data [ 19 ] The Lexus data was more complex, and was stored on a computer system associated with an Electronic Data Recorder [“EDR”], which controls airbag operations. That data was presented in Voir Dire Exhibit # 1, a six page printout of the data including graphs and numeric values for the data. [ 20 ] The Crown sought admission of all information in Exhibit 1, but specifically the speeds and vehicle pedal records. B. Evidence on the Relevant Data [ 21 ] Mr.

Hans Timmer testified on behalf of the Crown to explain the manner in which the Lexus vehicle operation data was collected and its meaning. He is a field engineer for Toyota Canada. He explained that on Sept. 29, 2010 he retrieved data from the Lexus. Mr. Timmer explained that report indicated that events had been detected by the EDR, and this could be something like a rapid deceleration. The EDR had recorded two separate events, and in each instance the EDR retained certain information on the vehicle’s operation for a five second period prior to the event.

The data from events 1 and 2 make up the first two pages of Exhibit 1. [ 22 ] The EDR also recorded four “triggering events”: two “frontal crashes” and two “side crashes”, each of which had associated data printed on a separate sheet. However, the electronic record indicated that a total of eight “triggering events” had occurred, but the hardware only had the capacity to record four of those. [ 23 ] Mr. Timmer explained various aspects of the recorded data:

1. both the driver and passenger were wearing seatbelts during both events; 2. in event #1:

a) in the five seconds recorded, the Lexus’ recorded speed decreased from 122 to 90 kilometers per hour. However, that 122 value is potentially inaccurate because the vehicle’s electronic systems only record a maximum speed of 122 kilometers per hour;

b) the Lexus’ brake lights were engaged in the last two seconds of the record;

c) the gas pedal was initially at “mid throttle” but was no longer depressed one second after the record began;

d) the record includes “engine” values which indicate the car engine RPM decreased from 4800 to 2800 RPM (800 RPM is the idle speed); and 3. in event #2:

a) the recorded speed of the Lexus decreased from 122 to 56 kilometers per hour;

b) the Lexus’ brake lights were engaged in the last four seconds of the five second record;

c) the gas pedal was off throughout the record; and

d) the car engine RPM decreased from 4400 to 800 RPM. 4. the four “triggering events” recorded occurred in the sequence of front, driver’s side, driver’s side, and front. All four trigger events occurred in under a quarter of a second. [ 24 ] Mr.

Timmer acknowledged certain limitations on this data record: 1. he was not qualified to analyze the data; 2. he could not attest to the accuracy of that data; 3. the reports in Exhibit 1 were a synthesis of raw data, and he could not comment on the processes used by Toyota’s proprietary data processing software to produce these results; 4. the data recorded was not time and date stamped; 5. he could not conclude whether the four “triggering events” were associated with either event #1 or #2; 6. he could not provide data on the margin of error or measurement accuracy of the Toyota software on the Lexus; 7. the software tracks whether the transmission is set to manual or automatic, but not the particular gear setting when in manual mode; and 8. driving over a curb could trigger the vehicle frontal sensor.

IV. The Parties’ Positions A. The Defence [ 25 ] The Defence argued that unfair evidence must be excluded from trial: R v Harrer ; R v Buhay , 2003 SCC 30 , [2003] 1 SRC 631. The Crown failed to provide the context to permit the Defence to evaluate the vehicle data system data. Judgments including R v Broughton , [1996] AJ No 970 (QL) (Alta QB) , R v Curran , 2013 ABQB 194 , 565 AR 201, R v Gubbins , 2014 ABPC 195 , and R v Kilpatrick , 2013 ABQB 5 , 555 AR 110 illustrate how this rule should be approached in the present case.

Due to the late emergence of this factor, the only appropriate remedy was to exclude the data recorder evidence, rather than order further disclosure: R v Bjelland , 2009 SCC 38 , [2009] 2 SCR 651 . [ 26 ] The Defence also argued a Charter analysis is not necessary where evidence fails the general rule for admission and may be excluded on that basis. [ 27 ] The Defence position is that machine-based evidence has no associated presumption of regularity except where legislated by Parliament. In this case there was much reason to be skeptical of the data.

Evidence of this kind should be evaluated in isolation for whether or not it is plausible. Expert testimony to better qualify the data extracted from the Lexus and Mercedes was available, somewhere, however it was not put before the Court. [ 28 ] R v Gratton , 2003 ABQB 728 , 341 AR 201 is a close analogue. There, evidence of this kind was excluded because the software was proprietary, its method of operation was uncertain, and there was no explanation as to how data was collected and analyzed. The information obtained from the Mercedes and Lexus electronic systems should be excluded for the same reason. B. Crown

[29] The Crown acknowledged that the Court has both a common law and Charter right to exclude evidence that is unfair to anaccused. The Crown argued there is no need in this instance to take the Charter-based route; these are effectively the same analysis. [30] The Crown’s position was that there are no R v Stinchcombe, (SCC), [1991] 3 SCR 326, 120 AR 161 first-party materials that had not been disclosed that would assist in the

interpretation of the motor vehicle ECU and EDR data. It had nocontrol over evidence of that kind. In any case the onus was on Mr. Soni to locate and request production of that data via the procedure inR v O’Connor, (SCC), [1995] 4 SCR 411, 130 DLR (4th) 235. The correct approach in relation to technical operationand calibration data is provided by R v Cyrynowski (August 14, 2014), Edmonton 140195587S1 (Alta QB). [31] The Crown stressed that the data recovered from the Mercedes ECU and Lexus EDR is relevant to the unsafe manner inwhich the vehicles were driven.

Though there is no time and date or odometer-based evidence that links the recordings to the collisions,that is, nevertheless, the logical inference. [32] The correct approach to evaluate whether the motor vehicle recording data meets the threshold for admission is incombination with other corroborating evidence: R v U. (F.J.), (SCC), [1995] 3 SCR 764, 128 DLR (4th) 121. Here thatincludes the eyewitnesses and the crash reconstruction report. The test is to inquire whether the ECU and EDR data is unexpected, andthat is the basis to identify whether or not a machine is operating in a reliable manner.

The fact a device is potentially unreliable orunsuited for a task goes to the weight of that evidence, not whether that evidence is or is not admissible. [33] Here we have a technology that is analogous to speed detection systems such as radar guns, and blood alcohol concentrationtest apparatus. The fact that a device operator cannot explain its operation is irrelevant: R v Grainger (1958), (ON CA),[1958] OWN 311, 120 CCC 321 (Ont CA).

There is no need for an operator to understand the scientific and engineering principlesinvolved. [34] The Crown argued that mechanical devices are presumed to be prima facie accurate: R v Werenka (1981), AJ No 1038 (QL),11 MVR 280; R v AK, 2004 ABQB 875, 373 AR 141.

Operator evidence is potentially relevant to prove the reliability of these devices: Rv Redmond (1990), (ON CA), 37 OAC 133, 54 CCC (3d) 273 (Ont CA); R v Delorey, 2004 NSCA 95, 226 NSR(2d) 59. [35] That is what distinguishes this scenario from R v Gratton, as that case relied on expert evidence which inadequatelyreconstructed a crash scenario, rather than simple measurements. Further, the technical data in R v Gratton was not supported byeyewitness evidence. [36] In

summary, the Crown’s position was that all machines are reliable as they do not have a party interest. They followpredictable principles: “The laws of physics apply.” The evidence from the automobile electronic data systems is a kind of hearsay, and agood analogy is the business records rule that presumes materials of that kind are reliable. IV. Analysis [37] The admissibility of evidence is a balancing process. In R v J.-L.J., 2000 SCC 51 at para 47, [2000] 2 SCR 600, Binnie Jexpressed this simply as: ...

Thus the criteria for reception are relevance, reliability and necessity measured against the counterweights of consumption of time,prejudice and confusion ...

This is both a general principle of law (Morris v The Queen, (SCC), [1983] 2 SCR 190) but also is very important inevaluating novel kinds of evidence (R v J.-L.J.). [38] Reduced to its core, the issue for this voir dire is nothing more than an application of the general rule on the admission ofevidence: whether the probative value of the evidence outweighs its prejudicial effect (R v Wray, (SCC), [1971] SCR 272at 297, 11 DLR (3d) 673), unless that evidence is obtained in a manner inconsistent with an accused person’s Charter rights: R v Grant,2009 SCC 32 at para 80, [2009] 2 SCR 353. A.

Preliminary Issue – Corroborative Evidence [39] The Crown argued that I should not evaluate the Mercedes and Lexus electronic data system evidence in isolation, but insteadwithin a larger matrix of facts, including eyewitness testimony and the accident reconstruction report conclusions. The Defence arguedthat R v U. (F.J.) provides a more restricted approach to this kind of corroboration, as that may only occur where evidence isindependent and similarity is more than coincidence. [40] I did not think it was necessary to evaluate this factor for two reasons.

The first was that the accident reconstruction’sinformation pool included the ECU and EDR data. Using the report to test the reliability of that data would be an exercise in circularlogic. [41] As for other evidence, I rejected its relevance on the basis that it is improper for imprecise evidence to prove the accuracy ofmore specific data. This can be illustrated by an example. Mr. Guimares said the Lexus and Mercedes were going very fast, they wereprobably speeding. The Mercedes ECU recorded a speed of 105 kilometers per hour. The testimony of Mr.

Guimares tells us nothingabout whether the specific velocity recorded, 105 kilometers per hour, is accurate. However, Mr. Guimares can support the propositionthat the ECR’s data is accurate within a broader range, that the vehicle was moving fast and speeding. [42] This is a simple and logical conclusion, and true anytime one measures something in two ways, one of which is (allegedly)more accurate than the other. An observation that the less accurate method does not contradict the more accurate does not confirm the

latter is more precise. It only confirms the (allegedly) more accurate mechanism is not so miscalibrated that its malfunction cannot bedetected by the less accurate test. [43] I therefore concluded that any context offered by other evidence was irrelevant to evaluate the Mercedes ECU and Lexus EDRdata records. B. Probative Value [44] Evidence has probative value when that evidence:

a) relates to an issue at trial, and

b) is reliable and credible. [45] Credibility and reliability are therefore key to evaluate the weight of evidence, and therefore whether or not that evidenceshould be considered by the finder of fact. [46] Witness credibility and reliability involves two separate processes: 1. is the witness telling the truth? and 2. can the witness provide accurate evidence? [47] These two interrelated questions involve a wide variety of factors.

A useful restatement of this issue is provided Estay J inWhite v The King, (SCC), [1947] SCR 268 at 272 The issue of credibility is one of fact and cannot be determined by following a set of rules that it is suggested have the force of law and ...it cannot be supported upon the authorities.

Anglin J. (later Chief Justice) in speaking of credibility stated: by that I understand not merely the appreciation of the witnesses’ desire to be truthful but also of their opportunities of knowledge andpowers of observation, judgment and memory—in a word, the trustworthiness of their testimony, which may have depended very largelyon their demeanour in the witness box and their manner in giving evidence. Reymond v. Township of Bosanquet. The foregoing is a general statement and does not purport to be exhaustive.

Eminent judges have from time to time indicated certainguides that have been of the greatest assistance, but so far as I have been able to find there has never been an effort made to indicate allthe possible factors that might enter into the determination. It is a matter in which so many human characteristics, both the strong and theweak, must be taken into consideration. The general integrity and intelligence of the witness, his powers to observe, his capacity toremember and his accuracy in statement are important.

It is also important to determine whether he is honestly endeavouring to tell thetruth, whether he is sincere and frank or whether he is biassed, reticent and evasive. All these questions and others may be answered fromthe observation of the witness’ general conduct and demeanour in determining the question of credibility. [Emphasis added.] [48] In the current matter, the Crown submitted the data downloaded from the Mercedes and Lexus as a kind of “mechanicalwitness” testimony. No experts were called to interpret or explain this information. I disagree that this is a kind of hearsay evidence.

It isnot relayed from a third party to a witness. [49] Mr. Timmer and Mr. Bairos explained the procedure by which that data was retrieved and how it came into its documentaryform. Their other testimony was restricted to assist in interpreting these documents and explain aspects of the electronic systemsinvolved in the data collection and storage. Otherwise, these documents essentially spoke for themselves, and the mechanical systemsfrom which they were obtained. [50] Continuing the analogy of the Mercedes and Lexus electronic data systems as witnesses, there is little reason to doubt thatthey are credible.

These are machines; they have no motive to lie or mislead. Further, for these systems to perform their functions in auseful manner, they should be accurate. The engineers who created these systems logically had no other purpose. [51] So the issue was reliability. Can we believe what these devices told us? This is a difficult point, because both Mr. Timmer andMr. Bairos were very clear on the limitations of the data in question. The case law reviewed in R v Werenka stresses technical evidenceis reliable when operators are familiar with a technology and their conclusions are supported by control tests and data. Mr.

Timmer andMr. Barios could not explain the operation of the respective electronic data sensor and analysis systems. The electronic data reports haveneither controls nor indicate a normal range of variation. [52] In the case of the Mercedes ECU evidence, that data was being used for a purpose other than why it was recorded. The basisby which that data was collected, analyzed, and recorded is unclear. It was not linked by a ‘time and date’ stamp to the events of April23, 2010. The best possible potential link between the crash and the recorded “fault code” was that that fault occurred at 79,050kilometers.

A kilometer is a substantial distance in a matter such as this. Further, I received no evidence on where that 79,050 kilometerdistance potentially fell in the vehicle’s operating history. [53] The “fault code” may have occurred prior to, during, or after the crash. Worse, post-crash events, such as the vehicle beingtowed, may result in false data recordings. [54] The situation was similar for data recovered from the Lexus. Again, the mechanisms to obtain, interpret, and record the data

were unknown, as was the reliability of that data. The data was not linked to a specific time and place. [ 55 ] Reliable scientific and technical evidence has a number of key attributes: it is tested, subject to peer review, has a known or understood potential rate of error, and involves theories and principles that are generally accepted: R v J.-L.J. , at paras 33-35 . I do not agree with the Crown’s submission that the fact a device is potentially unreliable or ill-suited is irrelevant to whether the evidence is or is not admissible.

The reliability (or unreliability) of a technology is both a factor to determine if that evidence is reliable enough to be presented to the finder of fact, and also relevant to the finder of fact to evaluate what weight to place on that evidence. [ 56 ] Returning to the characteristics of a reliable witness as outlined by Justice Estay, I do not know: 1. the accuracy of the sensors used to collect this data; 2. the ability of the electronic systems to accurately evaluate that data; and 3. the reliability at which that data is retained and may be extracted. [ 57 ] The witnesses who commented on this data were able to identify certain limitations.

The Lexus EDR cannot record speeds in excess of 122 kilometers per hour. Are there other analogous limitations? Mr. Timmer and Mr. Bairos were explicit – they did not know. [ 58 ] Both data sets included velocities. How was that obtained? Was that information inferred from the rotation of tires or from some other source? Did that value take into account skidding? Did these vehicles have systems that adjust for loss of traction? I do not know. [ 59 ] Did the engineers and software programmers who created these systems aim for extreme accuracy, or instead focus on data that is ‘good enough’ for their purposes.

That is particularly relevant for the Mercedes, where the data was not collected for anything other than vehicle diagnostic purposes. [ 60 ] What was the range of error on these data points? Did the vehicle sensors lose accuracy over time if they were not calibrated? How often did they malfunction? Did a malfunction result in no data, or an anomalous value? Would a collision lead a sensor to provide a spurious result?

Again, I do not know. [ 61 ] The net result was that I had two honest, mechanical witnesses, but I could not tell whether: 1. their vision was clear or blurred; 2. they remembered what they saw; and 3. they were relating their observations of the April 23, 2010 collision, or something else entirely. [ 62 ] The Crown suggested the appropriate approach was analogous to devices such as photo radar systems, where a user does not require a technical understanding of a device for its data to be accepted as face value.

That, however, is the case because either those devices have been established in case law as reliable or they have been deemed reliable by legislation. However, the situation here is in contrast to the circumstantial evidence available to Slatter J (as he then was) in R v AK , where a technology was commonplace and well established (para 4), and the operation of the device was supported by a combination of procedures, controls, and fail-safe mechanisms (paras 13-14). That is not the case here.

I have no ‘control’ data or calibration evidence. [ 63 ] Another proposition advanced by the Crown was the presumption of mechanical regularity – that machines generally are accurate. That is a dangerous presumption on several bases. A clock may provide an entirely accurate time – for a different time zone. A machine may provide a very accurate measurement of a condition – and then have that measurement distorted as the machine is damaged. [ 64 ] The Crown argued that this data is logically linked to the crash. It was the last data collected.

However, that proposition is countered by the evidence of persons best positioned to evaluate that proposition. Mr. Timmer and Mr. Barios were explicit that other possibilities exist. [ 65 ] All these defects could, potentially, be addressed. I have little doubt that it is possible to establish the reliability of certain aspects of the data recovered from the two vehicles. However, that evidentiary foundation was not developed before me.

I have little doubt that there is, for example, information concerning the accuracy and operation of sensors that record the motor vehicle velocity, and that an appropriate expert or documentary information source could provide a better understanding of whether these values are or are not accurate. [ 66 ] All these factors compromise the reliability of the ECU and EDR evidence recovered from the Mercedes and the Lexus.

This data is clearly potentially relevant, as the motor vehicle velocities, if accurately measured and relating to the period of the accident, would suggest both these vehicles were travelling at a high velocity. That is consistent with the alleged criminal misconduct. But that evidence is worth very little if it is not accurate or relates to some unrelated period or events. [ 67 ] Justice Watson, in R v Gratton , 2003 ABQB 728 at paras 119-125 , 341 AR 201, evaluated the evidence of a Crown expert witness who commented on motor vehicle electronic data information he had downloaded from a crashed vehicle.

There were two issues: the expert had no knowledge of the basis for that information, and could not provide evidence to establish the reliability of the recorded data and downstream analyses. The result (para 123): ... placed that evidence in a category not much different from that of a person who was able to pull a book off a high shelf and read out pages from it. That would not make the contents of the pages admissible, particularly if the pages contained opinion evidence in a science not known or understood by the reader ... [Citation omitted.]

[68] The scenario before me was similar, except that no expert was involved. Instead it is I who was asked to take down that bookand attempt to augur its import. This is not an occasion for judicial notice, given the variables and the technical subject matter involved. C. Prejudicial Effect [69] The prejudicial effect of evidence flows from its capacity to mislead a finder of fact. Prejudice can emerge from a number ofprocesses. One is “moral prejudice”, that an accused is the kind of person who engages in the alleged illegal conduct.

For example,evidence of misconduct or bad character outside the offence situation being tried is a well-recognized example of potentially prejudicialeffect. That is why trial courts should presumptively exclude “similar fact” evidence: R v Handy, 2002 SCC 56 at para 64, [2002] 2 SCR908. [70] A second form is “reasoning prejudice”. This is the risk that evidence distorts the analytical process for the finder of fact. Thiscan result from a complex fact scenario, or evidence that may tend to lead the finder of fact to erroneous decisions.

Data that ispotentially decisive to the issue of guilt or innocence, which addresses “the ultimate issue”, is subject to particular scrutiny due to itspotential prejudicial effect: R v J.-L.J., at para 37; R v Mohan, (SCC), [1994] 2 SCR 9 at 25, 114 DLR (4th) 419 [71] A well-established basis for reasoning prejudice is novel technology or scientific evidence. Moldaver J (as he then was), in Rv Melaragni (1992), (ON SC), 73 CCC (3d) 348 at para 16, 16 WCB (2d) 329, identified eight factors that arerelevant when considering the admissibility of new kinds of technical or scientific evidence. One was: ...

Is the jury likely to be overwhelmed by the "mystic infallibility" of the evidence, or will the jury be able to keep an open mind andobjectively assess the worth of the evidence? ... [72] This factor was highlighted by Justice Sopinka in R v Mohan where he observed: There is a danger that expert evidence will be misused and will distort the fact-finding process. Dressed up in scientific language whichthe jury does not easily understand and submitted through a witness of impressive antecedents, this evidence is apt to be accepted by thejury as being virtually infallible and as having more weight than it deserves.

As La Forest J. stated in R. v. Béland, (SCC), [1987] 2 S.C.R. 398, at p. 434, with respect to the evidence of the results of a polygraph tendered by the accused, such evidenceshould not be admitted by reason of "human fallibility in assessing the proper weight to be given to evidence cloaked under the mystiqueof science". [73] Obviously, this “mystique of science or technology” was in play here. However, there another kind of “mystique” in theevidence in question: the “mystique of precision”.

Technology often provides what appears to be very accurate, specific information.This is particularly true of digital rather than analog technologies. The data in Exhibit 1 is seductive in that sense because it appearsprecise. [74] For example, the graphs for the two events recorded by the Lexus EDR provide in themselves a kind of narrative. The readersees how different data correlates, with a foot moving from a throttle to brake pedal.

However, closer examination shows how noteverything is as it seems. [75] The engine RPM values from the Lexus EDR in event #1 have values at five to zero seconds of 4800, 4400, 4400, 4000,2800, 2800. Read in isolation it would seem this data is being recorded as a precise value – for example, five seconds before the event itwould seem the sensor and onboard software system counted 4800 individual rotations over the span of one second. But viewed as a set,the data reveals that is not the case. All values are multiple of 400. The same is true for event #2.

That strongly implies the sensor and/orcomputer software is tracking this variable in a much ‘coarser’ form. That may be because the EDR rounds RPM’s to the nearestmultiple of 400, or that the sensor only counts revolutions in ‘400s’. We do not know. [76] Regardless, a finder of fact who evaluates data like this, which appears precise, is misled by the manner in which the data ispresented – when that information is presented without an adequate context on that data’s reliability. Somewhat ironically the Crown’sargument made that very point.

It is true that the laws of physics do permit very exact calculation of movement, but the result of thosecalculations is only as accurate as the means that measures the relevant variables. [77] This is a powerful prejudicial effect. Lay persons – and in this sense I mean both judges and jurors – are not trained to evaluatedata for precision, nor will they usually be equipped to conduct a critical evaluation of the reliability and variability of machine-generated data. I believe it is fair to observe that persons will tend to trust a ‘hard number’, unless they are warned of the risks involvedin that.

After all, “the numbers don’t lie” (except when they mislead). [78] As McLachlin J (as she then was) observed in R v Harrer at para 46, the potential for evidence to mislead makes it unfair toadmit: Evidence may render a trial unfair for a variety of reasons. ... Its potential for misleading the trier of fact may outweigh such minimalvalue it might possess. ... [79] In this case I was warned that these apparently precise values should be viewed with suspicion. The very individuals who usethis information on a professional basis expressed caution in how this data was obtained and its accuracy. Mr. Barios and Mr.

Timmerwere little better off than myself. They too observed numbers generated by a black box of largely unknown content or operation, andwere up front and admitted that was exactly the extent of their knowledge. D. Probative vs Prejudicial Effect [80] My next step was to evaluate the factors that favour this evidence be admitted or excluded. This is a balancing exercise; even

highly prejudicial evidence may be admitted if it has sufficient probative value: R v Shearing , 2002 SCC 58 , [2002] 3 SCR 33. [ 81 ] I concluded that evidence of this kind has a significant prejudicial effect but, even more importantly, it is potentially highly unreliable. The vehicle operation data was not linked with confidence to a specific time or event. The degree of accuracy of those measurements was not known. [ 82 ] To be clear, I did not conclude this evidence is highly unreliable, but potentially highly unreliable. The problem here is context.

It may very well be that evidence of this kind is very important in other cases, but it needs to be accompanied by an adequate foundation, and here that was absent. Once more turning anthropomorphic, automobile data systems of this kind may prove to be valuable witnesses both for the Crown and an accused, once we have a way to test their memory and determine the acuity of their senses. E.

Charter , s 7 Disclosure and Hearsay Issues [ 83 ] The Defence Charter notice also included the issue of whether the Crown had made adequate disclosure to permit the Defence to “test and vet the reliability of the black box information”. This then allegedly subverted Mr.

Soni’s ability to mount a full answer and defence, and breached his Charter , s 7 rights. [ 84 ] A significant portion of the Crown and Defence argument related to whether or not this additional technical evidence is a kind the Crown is obliged to obtain and disclose, or instead information held by a third party, and therefore may be produced under R v O’Connor . [ 85 ] Another issue was whether this evidence should be addressed via the principles for the admission of hearsay evidence, as reviewed in R v Khelawon , 2006 SCC 57 , [2006] 2 SCR 787.

I saw no reason to resort to principles developed for a special form of evidence, when an adequate answer is provided by the general rule on the admission of evidence ( Morris v The Queen ), which in any case provides the basis for more specific rules and exceptions in the law of evidence. [ 86 ] I will not comment further on these arguments, given they are irrelevant in light of my conclusion that the Mercedes and Lexus vehicle data evidence should be excluded as not having met the threshold for admissibility. V.

Conclusion [ 87 ] I therefore excluded the evidence in Exhibits 1 and 4 as inadmissible in this trial. Heard on the 27-29 th days of October, 2014. Dated at the City of Edmonton, Alberta this 18 th day of November, 2014. D.L. Shelley J.C.Q.B.A. Appearances: J.P. Quenneville for the Crown B. Beresh S. Straub (Student-at-Law) for the Accused