On October 23, 2019, Gorski Consulting conducted testing on Highbury Ave between Hamilton Road and Highway 401 in London, Ontario to document the road’s surface conditions. This was done in a manner that has been discussed numerous times on the Gorski Consulting website. It involved the attachment of an iPhone to the structure of a 2007 Buick Allure test vehicle. An App on the iPhone was used to document the longitudinal and lateral motion of the vehicle. Video cameras documented the position of the vehicle along the road including other factors such as the vehicle speed. This article will provide a general description of the testing site, numerical results from the testing and finally a discussion of how this testing relates to results from other similar highways in Southern Ontario.
Highbury Ave is the only, four-lane, controlled-access, freeway located in the City of London, Ontario which has a population of about 390,000. This expressway is only about 5 kilometres long. To the south it connects with Highway 401 (MacDonald Cartier Freeway) which is the highest-volume freeway in Canada, stretching from Windsor to border of the Canadian Province of Quebec. To the north the Highbury Ave expressway terminates at Hamilton Road, which is an old arterial roadway that crosses at a diagonal along the south-east quadrant of London. North of Hamilton Road Highbury is no longer a controlled access freeway but a four-lane arterial.
Highbury Ave contains two interchanges, one at Commissioners Road and another at Bradley Ave. It contains some sections of surface that are an asphalt pavement while in other sections it contains a concrete surface.
Several older features of the Highbury Ave make it less safe than other similar, high-speed freeways. One sub-standard feature involves the lack of additional surface beyond the painted, yellow, edge line, as shown in the example photo below. Several decades ago Highway 401 contained a similar lacking of surface width and this resulted in many loss-of-control collisions as vehicles drifted off the surface edge. Almost all freeways of the current age contain some additional surface width including rumble strips that warn drivers when their vehicle wanders too close to the surface edge.
Tall, non-native grasses have also begun to grow in some sections of the median of Highbury Ave. While some forms of vegetation can be helpful in decelerating a vehicle that has entered a median, not all vegetation is helpful. In this case the tall grasses provide minimal deceleration while also blocking the view of drivers across the median. Vision across a median can provide an additional second or more of warning allowing a driver to detect an opposing vehicle that has entered the median and may be approaching into impact. A second or two of additional warning can be an important difference in avoiding a collision or reducing its severity.
These are some examples of deficiencies that plague many old freeways that have not been upgraded. The City of London is expecting to conduct a re-surfacing of Highbury Ave, likely in 2021, however it is unclear what corrections will be made to improve its substandard conditions beyond its surface.
The need for an improved surface is exemplified in the results of the Gorski Consulting surface testing that was conducted on October 23, 2019.
The Testing Results
Five test runs were conducted on Highbury Ave on October 23, 2019. Each of the five tests commenced from the intersection at Hamilton Road and progressed southward past the Highway 401 interchange. Then the test vehicle was turned around and the testing was continued northward back to the Hamilton Road intersection. The intention was to conduct the five tests at increasing speeds, from 80 to 120 km/h. Unfortunately, due to the traffic volume, interference by traffic prevented a steady speed and in many instances the test vehicle’s position had to be changed from the right lane to the left lane and back again. These changes in speed and lane position had some effects on the data making it more difficult to compare the results from one test to the other. Never-the-less some interesting results were obtained. The following five figures provide the results from the five tests.
In the past we have attempted to make it easier for readers to differentiate between “good” and “bad” road surfaces by colour coding the values. Thus green coloured values, below 0.0200 indicate good road surfaces. Black coloured values, between 0.0200 and 0.0500, indicate acceptable surfaces that will likely contain local problems. Red coloured values, above 0.0500, indicate there are likely major problems with the road surface. An exception has been made in the above table to reflect the observation that it is important for high speed expressways to contain a higher level of service, less vehicle motion and therefore lower values of rotation. Thus in the above table several values have been coded in red wherever they rise substantially above the norm for what would be expected for expressways. This is to note that any expressway with a value greater than 0.0500 must be considered of greater concern than a similar value on a lower speed road with less traffic volume.
Without some background to the meaning of the data the significance of the posted results can be lost. Yet it is a challenge to review the background without going into a long and detailed discussion. So, the following will be an abbreviated background which will likely require readers to look at some of the previously posted articles dealing with the Road Data datafile.
In brief, the columns in the above figures labelled “Lateral Rotation” and “Longitudinal Rotation” provide an indication of how the body of the test vehicle was moved, bounced or rotated as a result of its travel over the road surface for the time period of 30 seconds. If the test vehicle was travelling at 90 km/h that would translate to 25 metres every second. So in 30 seconds the vehicle would travel about 750 metres or 3/4 of a kilometre. So the posted number of 0.0400 provides the average rate of rotation of the vehicle body over that time and distance. The number is displayed in radians per second. One radian is equal to 57.3 degrees.
Let us look at an example where the lateral rotation is noted as 0.0400. If we multiply 0.0400 by 57.3 we obtain 2.29 degrees. So this value says that, during the noted time segment of 30 seconds, the average deviation, from a level, non-rotating position, of the body of the test vehicle was 0.0400 radians or 2.29 degrees per second. Lateral rotation refers to the motion that occurs if we were to grab a hold of the door frame and began rocking the vehicle back and forth sideways. This motion is referred to “rotation about the longitudinal axis” of the vehicle because the longitudinal axis is the line that passes through the centre of the vehicle from the front license plate to the rear license plate. This sounds odd because we are talking about something “longitudinal” when we are referring to a lateral motion. But this is the technical description of what we mean by lateral rotation. Now, because we are talking about an “average”, or standard deviation, this means that individual deviations within that average could be quite different from that average. Thus, for example, if our test vehicle runs over a length of 1 metre of uneven surface this might cause a major upheaval in the vehicle’s motion during that short time period. And we would not detect that short but huge spike unless we looked more closely at the graphing of the rotation.
An example of this is shown below. This figure shows the Longitudinal (blue) and Lateral (Red) Rotations of the test vehicle in Run #1 as the vehicle was northbound and passed by the Commissioners Road overpass while travelling in the right lane. Most of the data is clustered within the range of 0.1000 to -0.1000 radians per second yet we can observe several large “spikes” in the Lateral Rotation. At least one of the spikes in the middle of the graph rises above 0.5000 radians per second.
It would be of interest to find out what specific portions of the road surface caused these spikes and this can be done with considerable accuracy because of the multiple video cameras that are attached to the test vehicle and these cameras are synchronized to the iPhone App which senses these motions.
Yet it needs to be kept in mind that the spikes occur over a very short time frame of just fractions of a second and that matters. If the motion occurs for just fraction of a second then although the rate of rotation might be very high it does not correspond to a large rotation. We would be more interested in those spikes where there are high rates of rotation but also when they exist for several samples in succession. This would mean that not only did we have a right rate of rotation but the longer time of that high rate means that the vehicle’s body actually rotated to a greater angle.
We have also mentioned in previous articles that we acknowledge use of the 2007 Buick Allure test vehicle means that the results may only be valid for that test vehicle. In other words the use of a different vehicle may result in different data and that difference could be important. There are many agencies that use specialized equipment to plot the smoothness of a road surface and because this equipment is somewhat standardized the data is comparable from one dataset to the next. While this is useful for agencies that need to evaluate matters such as wear of a surface and timing of surface maintenance those are not the same needs as ours. Our interest is in documenting how road surfaces affect the motion and therefore the stability of a typical vehicle that drives on the surface. We only need to know how our road data compares to other road data from other roads. While it may be interesting to compare the data obtained using a different vehicle that is not essential for our purposes.
A substantial amount of data has now been obtained from a variety of testing over the past 5 to 6 years. The Road Data datafile on this Gorski Consulting website now contains tested roadways from many parts of parts of Southwestern Ontario and several counties. It also contains data from expressways that are similar to Highbury Ave. We can now look at the data from several of these expressways and see how Highbury Ave compares. The table below provides a summary of testing that has been conducted just this year on these expressways.
This table enables a general comparison of the road surface conditions of these major expressways in relation to each other. It can be seen that, overall, the Lincoln Alexander Parkway in Hamilton and Highbury Ave contain the worst road surface conditions.