Research and Development of High Intelligent Robot

China has fundamentally entered the "fourth consumer society", economic growth has entered the stage of high-quality development, and national consumption has paid more attention to quality. Affected by this, domestic travel consumption is also facing a major change in quality promotion, people's demand for high-quality travel services will rise rapidly, travel agencies, scenic spots, hotels, airports, railway stations and other related industries and some upgrading services are imperative.

AoBo's welcome robot is also suitable for tourist attractions and as an intelligent service robot in the hotel industry. It can serve as a welcome, explanation, guide and other missions. The professional, interesting and technological image makes the guests relaxed and happy in the whole viewing process, enhances the sense of tourism experience and improves the level and connotation of the exhibition hall. It can also make the guests who come to visit the exhibition hall focus their attention on the products or services that need to be publicized in the exhibition hall, and better promote the brand publicity and outstanding reputation of the exhibition hall.

Recently, the Institute of Microelectronics team of the Chinese Academy of Sciences and Professor Liu Qi team of Fudan University have made great progress in the research of multimodal neural shape perception.

Development of Highly Intelligent Robots < o: p >

Traditional semiconductor technology also faces challenges in device integration and circuit complexity in building a multimodal perception system in the pulse domain. Therefore, there is an urgent need to develop a more efficient hardware plan for multimodal blending perception. Biological perception system has the characteristics of parallel and distributed sensory information processing, low energy consumption and high fault tolerance, which shows the important potential to overcome the traditional dilemma.

Development of Highly Intelligent Robots < o: p >

In addition, by decoupling the frequency and amplitude of the output pulse, the researchers can also obtain independent pressure and temperature information from the blending signal, which supports the fidelity and multi-modal perception ability of neurons about single-modal information. The team further combined the MFSN array with the spiking neural network to construct an artificial multimodal perception system, which successfully imitated the perception of multimodal information (temperature and pressure) and the classification ability of multimodal objects (I. E. Objects of different temperatures, weights and shapes) in the human somatosensory system.

The above effects will help to further build an efficient multi-modal pulse perception system in the future, and provide new ideas for the development of highly intelligent robot skills.