To date, early warning signs of acute decompensation of heart (ADHF) are not detected and medical treatment is adjusted only to the apparition of symptoms leading to hospitalization. 40% of patients hospitalized twice in a year for ADHF die in the following year. Existing devices - implanted or not - allow to acquire relevant physiological parameter(s) whose analysis can lead to early detection. However, these devices do not respond efficiently to economic medical issues or require complex implantation. Some of these devices only consider a single biomarker whereas the HF syndrome is mutliparametric. Some devices are comparable to our proposal in terms of parameters taken into account, but they are only relevant for a small proportion of the patients undergoing heart failure, those equipped with cardio defibrillators. Over this project, our final aim is to develop a fully-implantable and fully-integrated device able to track various parameters involved in the ADHF process with the aim of providing early detection. Examples include heart rate (HRV) and respiratory rate variability, cardiac (S1, S3) and pulmonary heart sounds as well as patient activity and position. The originality of our approach lies in the choice of a gastric implantation site. This implantation site has never been used before and opens a promising exploratory field associated with new potentially relevant parameters in the follow-up of HF. The added value of the gastric implantation site also lies in the compatibility with an endoscopic route thus minimizing complications and time of hospitalization. This 3-year project is structured to validate a data extraction process and demonstrate, through a preclinical study, the ability to track specific ICID biomarkers through (1) a multimodal device implanted in the stomach of healthy or HF pigs in ambulatory conditions and through (2) an appropriate treatment of cardiac and respiratory signals. This project faces a certain number of technical and scientific challenges: denoising of signals, pre-clinical model of adapted HF. Each of these challenges is associated with well identified risks for which we already propose adapted fallback solutions. The consortium is composed of four complementary partners; 3 laboratories and a private company. The industrial partner will bring its knowledge in the fields of electronics and mechatronics and will ensure the valorization of the results. The hospital partner (LRB) will bring its expertise to ensure the relevance of selected pre-clinical models and selected pathophysiological parameters. Finally, the scientific partners (TIMC, LTSI) will bring their skills in the field of cardio-respiratory parameters monitoring techniques and associated information processing methods.