Computer Assisted Map Design for Visually Impaired People – ACTIVmap

The project is based on an interdisciplinary approach combining geomatics, multimodal interaction, and digital accessibility.

* Needs analysis and state of the art: a review of best practices was conducted, synthesizing more than twenty years of research on tactile map design. This step made it possible to establish guidelines for standardizing representations (Wabiński et al., 2022).

* Semi-automatic design: ACTIVmap develops pipelines to automatically generate thematic tactile maps or representations of intersections from open data (OpenStreetMap). These processes are based on cartographic simplification, automatic inference of relevant content, and the production of files ready for 3D printing or embossing.

* Multimodality and automatic description: prototypes produce textual and audio descriptions of urban elements, particularly pedestrian crossings (Kalsron et al., 2023). The approach combines cartographic vision and artificial intelligence to enrich non-visual navigation.

* Interaction and prototyping: an interactive Web GIS was designed to generate personalized audio-tactile maps. This system allows users to adjust the scale, the type of information displayed, and the access mode (tactile, audio, or hybrid).

* Participatory evaluation: the prototypes are tested with visually impaired people, in partnership with specialized associations and companies (FeelObject), in order to validate their usefulness and ergonomics.

This progressive approach aims to ensure scientific robustness while producing concrete and transferable solutions.

The project has produced several major scientific and technological results:

* Methodological framework: publication of standardized recommendations for the design of tactile maps, which are now the benchmark in the community (Wabiński et al., 2022).

* Technical prototypes:

* a semi-automated pipeline for generating tactile maps of intersections from open data;

* a Web GIS for producing customizable audio-tactile maps;

* inference algorithms for automatically adapting the scale and cartographic content.

* Multimodal innovations: systems that automatically generate textual and audio descriptions of intersections from OpenStreetMap to complement tactile representations.

* Scientific output: more than a dozen publications in international journals and conferences, covering tactile cartography, cartographic generalization, AI, and accessibility.

* Open dissemination: the prototypes and tools developed are made available under a free license to encourage their reuse and improvement by the scientific and associative community.

These results contribute to a solid methodological and technical foundation for cartographic accessibility, combining academic rigor with the practical availability of open solutions.

ACTIVmap opens up several avenues for research and application:

* Pipeline improvement: towards complete and adaptable automation of map production, with user-adjustable parameters.

* Extension of interaction modes: integration of richer haptic feedback, and exploration via mobile terminals or interactive devices.

* Expanded evaluation: implementation of broader user testing, in collaboration with associations for the visually impaired, to validate robustness and usability.

* Standardization and interoperability: contribution to international standards for tactile and audio-tactile maps, in order to promote consistent and replicable practices.

* Open ecosystem: consolidation of open source prototypes into a sustainable platform, integrating design tools, databases, and accessible services to encourage reuse in education, mobility, and urban inclusion.

ACTIVmap is thus part of an open science dynamic, where results are shared to maximize their social and scientific impact.

The production of maps accessible to people with visual impairments (PVIs) is nowadays a predominantly traditional practice, mastered by professionals in document adaptation, and essentially available for locomotion instructors.
Access to such representations for people with visual disabilities is an essential element, both for the understanding of the surrounding space, but also for mobility assistance. The generalization of the production of such maps is therefore an essential issue for the autonomy of these people. Access to these media should also be facilitated by increased interaction and sound information.
These needs are nowadays even more complex to satisfy as each user has specific needs in terms of representation (different tactile abilities, ability to represent space), and in terms of space to represent (maps at the scale of the district, plans and diagrams of complex intersections for the crossing, public transport network, etc.).
Scientific advances in geomatics (LASTIG) and geometry (LIMOS), combined with the availability of collaborative geographic data (OpenStreetMap type), new on-demand manufacturing techniques (3D printing) and interactive approaches (IRIT) offer original possibilities for designing such maps.
In this project, we will work closely with the various users of these maps: PVIs, orientation and mobility instructors, tactile document makers, special education teachers, and families.
We will work on collecting the different needs, focusing on two scales where the need is identified and well defined: the neighbourhood scale, and the crossroads scale.
We will then explore different ways of processing geographical data in order to produce representations adapted to these needs: production of maps by adapting the geomatic processing workflow (data selection and fusion, generalization, simplification and stylisation), production of abstract schemas and text descriptions, integration of geometric information from orthophotos. We will work to develop methods that integrate the adaptation to the diversity of disabilities and uses of these geographical representations.
These representations will be used when designing interactive devices adapted to the different uses and needs and capacities of the different users: fixed support for consultation and preliminary exploration, or mobile support for on-site consultation. In both cases, these devices will be interactive and audio, allowing the user to appropriate the available information. The automatic generation of this geographical information will make it possible to integrate data from geolocated databases or web services (information on the opening of public places, information on ongoing cultural events, for example) into the devices.
An open source software suite integrating the various research prototypes will be consolidated during the project in order to provide users with a simple way to manufacture the data needed to manufacture such maps.
While these explorations are in line with the scientific explorations of the three university partners, they are also in line with the activities of the industrial partner (FeelObject), which recently developed a peripheral device (Virtuoz) for PVIs, using 3D printing, and allowing interactive exploration of building plans. The application of this device for external use, the integration of the automatic processing explored during this project, and the adaptation of the initial product around interaction issues will consolidate the dissemination of the project's progress.