"The CyborgLOC project aims at the pre-industrial production of a multi-sensor adaptive solution for nomadic Indoor Outdoor geolocation. CyborgLOC relies on: a) the state of the art achieved by the IFSTTAR laboratory in inertial navigation (between 0.35% and 2% deviation over one kilometer), b) on Deep Learning methods And Big Data to integrate the real-time recognition of human body movements and subtract them in the trajectory calculations of the barycenter, c) on the expertise of the miniaturized microsystems of geo-location with energy harvesting of SGME as well as its mastery of sensors D) on the first prototypes of SGME for an Indoor / Outdoor Geolocation system based on the work of fusion and scheduling of data and calculations. The consortium brings together four major complementary domains to solve the challenges of the challenge: a) Inertial navigation, porting and transformation on the CyborgLOC platform of IFSTTAR's very advanced algorithms, b) Robotics, integration of algorithms and knowledge of Elter for Scheduling and behavior responsive to the environment and situation (including human body movements), c) Microsystems, miniaturization and integration of electronic microsystems of SGME (Bageo), with a search for energy saving Up to energy harvesting. These four major domains finally meet around a common theme: an adaptive geolocation system, focusing on movement and environment recognition based on deep learning algorithms for scheduling. The 1st CyborgLOC Init prototype has limited but still superior performance to the state of the art. The second prototype CyborgLOC Expert has all relevant sensors, distributed on the body. It uses distributed sensors for the recognition and classification of movements of the human body. The whole scientific method with the relevant algorithms are implemented. Multi-level merging and scheduling are effective. 80% of the knowledge learned and modeled in the successive test benches is integrated with CyborgLOC Expert. The stereo video sensor is implemented. The initial configuration phase of the competition takes advantage of a reinforced learning procedure dedicated to the wearer. The 3rd prototype ""CyborgLOC Tuning"" takes all the elements of feedback and enters a phase of tuning algorithms, performance and acuity. Several levers are used: (a) Classification and speed of recognition of improved body movements, (b) Improved parameterization and configuration of algorithms, (c) Increased calculation accuracy by increases in energy performance, (d) Scheduling improvement, With the environment more reactive, shorter transition phases, f) Personalization phase by enhanced learning on the individual improved. This phase also requires improving the measurement bench, improving analytical tools, and possibly switching to algorithms that are even more adaptable, such as the dynamic integration of new classes in Deep Learning. The creation of an outdoor / indoor geolocation system without infrastructures has the effect of positioning this system at the front of the state of the art. The members of the consortium would benefit from a favorable position to quickly reach certain professional markets. Moreover, reaching such a level could allow the development of a declination of the device towards the general public by reworking aspects such as integration with the existing (smartphone, connected objects), price , Design and ergonomics."