The safe transportation of children on school buses is paramount. An important factor in that safety is the reaction of school buses to road surface defects. Are school buses prone to loss-of-control because of their inferior characteristics compared to other vehicles? Are they superior in their design and construction and therefore ideal for safely transporting children? Little or no discussion exists in the public domain regarding this important issue. Gorski Consulting is conducting testing of school buses to address this issue.

Two previous articles have recently been unloaded regarding this testing. In the first article posted on April 5, 2021, testing was reported using an 18-passenger school bus driven along a path including Wharncliffe Road in London, Ontario. In a second article, posted on April 13, 2021, the same school bus was driven along a path including Wellington Road. Each of the two testing sessions were performed on March 4, 2021.

The present article will report on testing performed with a full-size International school bus which was driven on White Oak, Southdale and Wonderland Roads in London, Ontario. For simplicity this path will be referred to as the “Southdale” path. This testing session was performed on March 25, 2021. Two additional testing sessions were performed with this school bus on March 25 and 26, 2021 and it is hoped that time will be available to report on these sessions in upcoming articles.

The photo below shows a view of the 2012 International school bus that was used in the testing on March 25 and 26, 2021.

The following three bar-charts will provide the results from the previously-mentioned two testing sessions of March 4, 2021 as well as the current testing session on Southdale Road of March 25, 2021.


Recall that what is being reported in the three charts is the extent of motion caused  by the bus travelling over the noted routes. The blue bars indicate the extent of Longitudinal Rotation or the extent that the bus has been “bounced” front and back. The red bars indicate the extent of Lateral Rotation or the extent that the bus has been “bounced” left or right, or sideways. Each of the columns represent a time of 30 seconds of motion. When the bars are extremely low then this indicates that the bus has come to a stop, for a traffic signal for example.

Because the reported motions can occur in opposite directions their magnitude could be cancelled out if the actual rate of motion was reported. So this problem has been alleviated by reporting the “standard deviation” of that motion. In other words we are not concerned with whether the vehicle’s front end has been lifted or dropped, or if the bus body has moved down on the left or up. We are simply interested in the magnitude of that motion. So by using the standard deviation we are reporting, on average, how quickly the motion occurred regardless of its direction. It seems reasonable that if the rate of vehicle motion is higher then this must mean that the vehicle’s motion was affected to a greater extent by the road surface and that this is undesirable.

We must also pay attention to the speed of the vehicle during the testing because that has an effect on the magnitude of the motion. So if the bus is stopped and idling there is very little motion of its body. And if it is travelling very quickly then its reaction might be greater to the contact of an irregular road surface. While this may appear to be a confound it is of minor importance when an analyst become familiar with the signature of the expected motion.

An important question that  will hopefully be answered with this school bus testing is whether our procedures can be reliable in providing a useful, objective reporting of possible safety concerns when a vehicle rides over a particular road surface. For several years testing was performed by Gorski Consulting over many roadways in southern Ontario using a 2007 Buick Allure passenger car. This “Road Data” is reported in the Road Data webpage of this Gorski Consulting website. That testing has shown that the procedures we used could reliably differentiate one surface from another. The school bus testing is adding another dimension to the issue: When a vehicle with a totally different body is used as the test vehicle can the Road Data still reliably differentiate between the characteristics of one road surface versus another? From the two preliminary tests with the GMC 18-passenger school bus on March 4, 2021, the answer would appear to be yes, the methodology appears to be reliable when the test vehicle is a small school bus. Of course we still have to examine further testing with the full size school bus.

An interesting result of the testing with the International, full-size school bus is that the overall motions (shown in the above bar chart) appear to less in magnitude and the Longitudinal Rotation appears to be less than the Lateral Rotation. Below is a summary of the results from each of the three testing sessions.

March 4, 2021 Wharncliffe Route: Longitudinal = 0.0192, Lateral = 0.0214

March 4, 2021 Wellington Route: Longitudinal = 0.0206, Lateral = 0.0201

Mar 25, 2012 Southdale Route: Longitudinal = 0.0091, Lateral = 0.0144

We can also compare the above data with the larger data contained in the Road Data webpage of this website. The City of London and Oxford County are two of the larger datasets and the average for all roads in each of those jurisdictions is summarized below.

City of London – All Roads: Longitudinal = 0.0270, Lateral = 0.0265

Oxford County – All Roads: Longitudinal = 0.0245, Lateral = 0.0214

Also, Hwy 401 has been noted to produce the lowest motion data indicating the highest quality of road surface. Some data from Hwy 401 are noted below.

Hwy 401 WB between London & Tilbury: Longitudinal = 0.0126, Lateral = 0.0096

Hwy 401 WB between Woodstock & London: Longitudinal = 0.0134, Lateral – 0.0092

A word of caution is worthwhile. The testing inside urban areas such as the City of London incorporates many instances where the vehicle comes to a complete stop. Also the urban data is collected from a lower speed than the rural data. So, if all roads were equal in character, one should expect that the data from the urban testing should contain lower motion data.

While the data from the International School Bus testing showed lower motion data it must be observed that, at the beginning of the testing session, the bus was stopped in the bus yard for 200 seconds (about 3.5 minutes) before the bus began its movement. Also the bus ignition was not activated until 150 seconds (2.5 seconds) had elapsed. So during this time when the bus stood still the motion was very low and this obviously lowered the overall average for that session. However the curious finding is that the overall Longitudinal motion was substantially lower than the Lateral motion. That finding does not seem to be confounded by the bus being stopped. At least there does not appear to be an obvious confound at this early stage of study. Given that a full-size school bus is very long (12.5 metres) compared to the other two test vehicles it is possible that this long length could reduce the forward-rearward bouncing of the bus and this could explain the lower Longitudinal Rotation values. However this is just a very preliminary thought and we will see how this unfolds upon exploring the remaining two testing sessions. We hope to report the two remaining testing sessions with the full size school bus in the near future.