Think remote diagnostics = IoT? Think again
Especially in trucking but also in a variety of other industrial sectors, many manufacturers have added telematics units to their vehicles in order to collect diagnostic codes and transmit them to cloud databases. Having done so, they’d like to think they have joined the latest technology mega trend—Internet of Things (IoT).
Unfortunately, they also quickly find that, while useful, these basic data gathering systems fail to deliver the business outcomes they are looking for (e.g., increased asset uptime, reduced service and warranty expense, etc.). This is because they are not yet fully exploiting the potential of IoT. Rather, they have only taken the initial steps required to build an IoT system; at best they have built an M2M system, says Dave McCarthy, director of Products at BSQUARE.
With remote diagnostic initiatives, manufacturers encounter a number of problems. First, the sheer volume of data they have to deal with quickly becomes overwhelming. Many manufacturers have expressed surprise at how chatty their trucks are.
True, most of this is informational or minor diagnostic codes, but the noise generated by this river of data makes it more difficult to detect truly anomalous conditions. For this reason, data reporting frequency and data granularity are often reduced in order to make it more manageable for human operators (and, not incidentally, make sure network and storage costs don’t balloon out of control).
But this directly and adversely impacts the ability of these systems to perform the basic function they were designed for—detect operational problems. In other words, they often cripple the system they built because it is overwhelming.
Second, most of these systems perform a very simple classification of error codes ranging in severity from minor (e.g., washer fluid is low) to major. What they are looking for are diagnostic codes indicating the need for servicing of the vehicle either immediately or after dropping the current load.
However, because of the relatively simplistic classification methods and lack of robust data analytics, many manufacturers still endure large numbers of false positives (the truck was sent in for servicing but resulted in no problem found) and false negatives (the truck should have been sent in for servicing but was not). This results in still-too-high servicing costs and still-too-low asset uptime.
The reason is that actual service-inducing events are often much more complex than a single diagnostic code can convey. Experience has shown that being able to look at historical data, along with surrounding and contextual data, in addition to data coming off the truck, often results in event profiles that are more of the form “if a, b, and c occur within 30 minutes of each other and d does not occur, then trigger a major event” than the more simplistic “if a then b” common to basic remote diagnostic systems.
IoT can address these issues by bringing data analytics, rule processing, and automated orchestration of business actions to what is basically an M2M construct. By doing so, manufacturers, having implemented true IoT systems, will be able to achieve the business outcomes that actually matter.
The author of this blog is Dave McCarthy, director of Products at BSQUARE.
Comment on this article below or via Twitter: @IoTNow_ OR @jcIoTnow