Every day, MiCa shares its routes with pedestrians, cyclists and other vehicles. These encounters are part of normal operation, and our job is to make sure MiCa keeps handling them consistently as the vehicle and its software develop. That is why testing is a continuous part of our work: we repeat the same situations, measure how MiCa responds and compare the results as the system develops.

From simulation to real-world validation

MiCa’s testing starts in simulation, where new software versions and driving scenarios can be tested and repeated efficiently. From there, development moves to our own test ground, where our team regularly tests software updates, improvements and the shuttle’s reactions with the physical vehicle.

Testing with Metrosert adds another layer: precise, repeatable real-world validation. Scenarios tested in simulation can be recreated with MiCa in the physical world, with defined positions and timing. This lets our engineers compare simulation results with the behaviour of the real vehicle and check how closely the simulator reflects what happens on the road.

From the test site to live traffic

In August, MiCa was tested at Metrosert’s proving ground in Paldiski, where controlled, repeatable scenarios allowed us to compare simulation results with the physical vehicle.

The tests in Tiskre build on that work. In November, selected scenarios moved from the proving ground onto a residential street in Tiskre, in Tallinn’s Haabersti district. The use of the street was coordinated in advance with the relevant road and city authorities. The scenarios stayed controlled and measurable, but MiCa also had to account for the surrounding traffic.

What we tested

Pedestrian at a crossing

‍A pedestrian soft dummy enters the crossing as MiCa approaches. It is one of the most common situations on a shared route: the shuttle has to detect the pedestrian and give way.

We started with the pedestrian scenario alone, then added other road users, from one vehicle to several in different positions around the crossing.

‍

Cyclist crossing

‍A cyclist soft dummy crosses MiCa’s path. A cyclist moves faster than a pedestrian and reaches the crossing sooner, which changes the timing of the interaction.

As with the pedestrian, we started with the cyclist alone before adding other traffic in different situations.

Testing with Metrosert in Tiskre proved a valuable experience and confirmed that MiCa behaved as expected in the tested scenarios, just as it had at the proving ground in Paldiski.

What we do with the data

Metrosert measures how and when MiCa responds in each run. Our engineers use this data to analyse the vehicle’s behaviour in detail, compare physical tests with simulation results and see how software changes affect its response.

Because the scenarios can be recreated precisely, selected tests can be repeated as the software develops. That gives us a consistent reference point across simulation, controlled testing and operation in real traffic.

About Metrosert

Metrosert is Estonia's National Metrology Institute and an Applied Research Center, maintaining the national measurement standards and developing new technologies with companies. Its Autonomous Vehicles Unit, one of the center's six research areas, operates a mobile field test laboratory for automated and remote-driven vehicles. The unit recreates traffic scenarios precisely and repeatably on its proving ground in Paldiski and on public roads. It also trains safety drivers and remote operators and helps companies prepare for pilot driving and type approval in Europe.

‍