Revolutionary solutions for man-machine interactions appeared during the last ten years. Tactile interfaces (smartphones, tablet PCs) and motion sensing and control (Wiimote, Wii Balance, KinectTM) are now available for general public while in industry, immersive environments and haptic devices boosted ICT productivity in design and training. However, no existing interface really allows natural hand motion control and dexterous interactions. Haptic interfaces limit the user dexterity while data-gloves are uncomfortable, difficult to calibrate and do not give any feedback to the user. The design of truly dexterous peripherals allowing natural and intuitive mono or bi-manual interactions with force feedback in virtual environments is indeed a real scientific challenge. The MANDARIN project proposes an innovative 4 points approach to answer this challenge: • To deal with the structural and functional complexity of the human hand, and with its extreme inter-individual variability, we will design accurate biomechanical models allowing to analyze the kinematic compatibility of dexterous haptic devices with the hand structure and movements. • These models will be used in combination with technological breakthroughs to design an innovative comfortable and high performance force feedback glove. It will be sufficiently simple, compact and light for an easy usage by non specialists and for long term mass production at a competitive cost. • Considering software, we will design novel multimodal 3D interaction techniques and metaphors allowing to deal with haptic gloves limitations (portability, under-actuation) and to assist the user in virtual reality applications requiring dexterous manipulation. • New physical simulation algorithms and dexterous interactions simulation software allowing to compute the finger forces and to communicate them to the haptic glove controller in interactive time will allow practical usage of the glove in virtual environments. These scientific advances will be evaluated with a representative industrial application which is not feasible with currently existing technologies (bi-manual manipulation of complex rigid objects and cables bundles). Being involved in the design of the prototype, HAPTION will be ready for the development of a commercial product which could be launched about one year after the end of the project. Industrial end-users continuously replace their physical mock-ups with digital ones and would be interested in using such high-tech products as much as possible for increased productivity allowing huge cost savings. Beyond industrial applications, the MANDARIN project aims at designing general solutions which can be used for any kind of application requiring dexterous manipulation of virtual objects, either in industry (digital mock-up, fitting, design reviews), digital arts (virtual museums), education and training or health (rehabilitation, assistance). These results could also be licensed to high-tech companies for consumer market applications (video games) like was the KinectTM for motion capture. These results will also constitute truly innovative tools for future researches in haptic perception, dexterous manipulation and interactions in virtual environments (neurosciences, psychology, physiology).