Principles for grouping interlock ACC, start, and test events into trips

5 min read

An interlock is an alcohol interlock that controls whether a vehicle can start based on the result of the driver's breath test. With the DATECH AF-i1, the test is taken on the handset and the control box connects to the vehicle to operate it, storing vehicle ACC, engine start, test result, and tamper-related events with date and time. ACC here is a record related to the vehicle's power state; it does not mean the same thing as an engine start record or a test result record.

Fleet managers or staff reviewing logs may try to work out "when did one trip happen" from these individual records. However, the list of events the device stores is not a ready-made list of trips. A trip segment should be treated as an analysis result, created by deciding which tests and start records to view as one group, without altering the original events.

First separate original events from analysis segments

The official AF-i1 information states that it stores data related to interlock use and that log events can be managed and downloaded with dedicated software. However, this information alone does not mean the software automatically provides a finished segment called "one trip." What names and order the ACC, engine start, and test result records come down in must be confirmed from the records you actually download and from the product documents.

So keep the event names, dates and times, and displayed results the device recorded exactly as they are in the downloaded original. On top of that, build the trip segments you create for your business purposes as a separate list. If you correct the original or merge several records into what looks like a new event, it becomes hard to trace later why you made that judgment.

When building trip segments, it is good to record three judgments separately. First, decide which events to treat as candidates for the start of a segment. Second, examine whether there is a basis for linking a preceding test result to that start record. Third, decide how far the previous segment extends when the next start or ACC record appears.

Do not link records automatically just because they were recorded close together in time. In particular, if there is only an ACC record and no engine start record, there is not enough basis to conclude that the vehicle was actually driven. Conversely, if there is an engine start record but it is unclear which test result it follows, it is accurate to leave the test link as "not confirmed."

What to judge What to look at in the records Conclusion to record in the analysis list
Segment start Whether there is an engine start event Start confirmed or start unconfirmed
Test link The preceding test result and time order Basis for link exists or link unclear
Segment boundary The next ACC or start event and any gaps in the records End estimated or end unconfirmed
Record reliability Possibility of omission or filtering, time order Whether further checking is needed

Dividing the conclusions as in this table makes clear what has been confirmed and what is missing, rather than using a catch-all phrase like "looks like a trip." For the same reason, do not fill in boundaries by assuming there are engine stop records, location data, or driver identification that the device does not provide.

Interpret repeated tests and starts separately

As a hypothetical example, suppose a vehicle's records show two consecutive test results followed by one engine start event. A short time later an ACC-related record appears, but there is no further start event. The facts to observe here are "two tests, one start, one ACC-related record," and you cannot immediately conclude that the two tests each represent separate trips.

Interpretation starts from the event definitions the device actually recorded and their time order. Whether the second test was a test required again before starting, or whether there was separate preparation after the first attempt ended, must be looked at together with the original event names and the operating situation. A possible action is to link both tests to the same start candidate, while recording in the analysis list which result was linked to the start and the basis for that judgment.

The completion criterion in this case is not deleting the two tests arbitrarily or confirming them as separate trips. It is enough that both original test records and the one start record can be traced, and that whether the link to the start is confirmed or on hold is marked. Record the date of any change to the judgment rules so that the same criteria can be applied when you process the same type of record later.

Do not fill unconfirmed segments with normal trips

If the download range was limited to part of the period or intermediate events are missing, there may be no pre-start test record or nothing that corresponds to the end of a trip. In such cases, rather than filling the empty segment with a normal test and normal trip, mark the state of the boundary, such as "start unconfirmed," "end unconfirmed," or "test link unclear." Separating the count of complete segments from the count of segments with unclear boundaries keeps the limits of the records from being hidden in the analysis results.

The AF-i1 guidance states that it can store about 500,000 test results and operating events, and it also describes inspection and calibration reminders. If you find a longer-than-expected gap or out-of-order entries while reviewing records, you can have an operating procedure that first compares the download range for that time with the device's inspection records. This is not to decide the cause of the gap but to sort out whether the records are ones to which the analysis rules can be applied.

Record the basis for judgments alongside the results

In the trip segment list, record the date and time and event names of the original events, the linked test results, the start and end judgments, and the reasons for anything unconfirmed. Even for the same vehicle, if the handset and control box configuration or the time of record collection changed, do not automatically stitch different sets of records together. The AF-i1 handset performs the tests and the control box acts as the central processing unit, so you need to confirm the actual configuration and collection range before judging whether records come from the same device.

A trip segment is not a fact the device left behind but an analysis rule for reading those facts. Only by preserving the start and end, the basis for test links, and the remaining gaps together can you explain later why ACC, start, and test events were grouped into one trip.

References

#AF-i1
#interlock event log
#trip sessions
#ACC and engine start
Considering our products?
We'll help you decide.