Imagry and Spanish bus manufacturer Unvi have integrated Imagry’s HD-mapless autonomous driving software into a Unvi-made battery-electric bus, with the first vehicle already en-route to a deployment site in Germany ahead of a planned Q3 launch. Israeli firm Ubsi is handling procurement, homologation and fleet operations as the integration partner between the two companies.

The vehicle in question is a 10.3-metre, M3-category e-bus, pre-certified for European Whole Vehicle Type Approval under the bloc’s GSR2 safety regulation. Imagry and UNVI say the software integration itself took just shy of three months, which is well below typical industry timelines. Neither company has disclosed the bus’s battery specifications or range, nor have they detailed the exact German deployment location.

It is also unknown whether the HD-mapless service will operate fully autonomous from launch or initially run with a safety driver present. With that said, previous autonomous bus launches—even those with more established technologies than Imagry’s—usually begin with a safety driver present, if for no reason other than reassuring riders.

It is certainly noteworthy that the service will go live without HD mapping, because conventional engineering wisdom actually favours the opposite approach for this kind of vehicle. Fixed bus routes are predictable and repeated, and that is precisely what HD maps exploit best: centimetre-level curb geometry that enables docking precision for wheelchair ramps, a pre-built world model that frees onboard compute for tracking pedestrians and traffic instead, and a stored spatial reference that keeps working even when heavy rain or snow obscures lane markings from cameras.

Because transit authorities only need to survey a fixed set of routes rather than an entire city, HD mapping is also comparatively cheap to maintain in this specific use case. Imagry’s mapless approach inverts that logic on several practical grounds that matter specifically to transit operators. Skipping the survey, point-cloud processing and validation an HD map requires is precisely why software integration took under three months rather than years, letting operators launch or extend a route in a matter of weeks. European roads also experience frequent roadworks, closures and detours that could strand a map-dependent bus when reality no longer matches the stored data.

This is the advantage of a mapless approach: it simply reads the road as it finds it. With that said, there is a real trade-off. HD mapping’s strongest specific advantage, guaranteed sub-inch docking against a pre-surveyed curb, becomes harder to replicate without stored measurements. To address this, Imagry is pairing its self-driving perception system with additional sensors trained specifically to detect platform edges in real time.

The launch lands mere weeks after UNECE adopted the first internationally harmonised regulatory framework for automated driving systems at SAE Level 4 and above, backed by the EU, US, China and Japan. The framework is not binding, and the Imagry-Unvi German deployment will initially run under existing national WVTA and GSR2 rules instead. The country’s broader push toward automation—distinct for its preference for public and commercial transportation over robotaxis—is driven in no small part by a worsening driver shortage. This is not unique to Germany, however: globally the transit industry is facing a 20% global shortage that is worsening by roughly a percentage point each year.

Unvi plans to extend the partnership to other vehicle configurations over time, and both companies see potential beyond conventional public transit routes in campuses, airports and hospitals. All of these are environments where a defined but changeable operating area could potentially suit a perception-based system better; one dependent on pre-surveyed HD maps that would need updating each time a route or building layout changes.