This paper discusses the dynamics of emotion recognition on faces, layering basic capabilities of an emotion sensor. It also introduces a model for the recognition of the overall conveyed emotion during a human-machine interaction, based on the emotional trajectory over an emotional surface.

Lower limb orthoses are often plagued by a high rejection rate, due both to the excessive effort demanded from their users and to the lack of aesthetics of the resulting gait. These factors are caused, to a great extent, by the fact that orthoses force the whole gait to be performed with the knee articulation in full extension. This work presents a light and compact device, with a low energy consumption, that can improve the performance and the gait aesthetics of knee-ankle-foot orthoses (known as KAFO). To do so, we explore the natural dynamics of the lower limb through a spring that is attached to a standard orthosis system. We also add control circuitry, a small electric motor, sensors and microprocessors to the standard orthosis, so that we can control the whole gait in a relatively natural manner. We have designed the system through a large set of simulations and have shown that the passive dynamics of the lower limb, driven by a burst of energy from a spring at the beginning of the swing phase, is sufficient to reduce the compensation mechanisms required during the gait with orthosis. Thus, this strategy is an alternative to existing solutions relying on functional electrical stimulation (FES), which suffers from limitations such as rapid muscle fatigue and difficult motion control. We describe the design of the device, the model adopted for the swing phase of the gait, the numerical simulations performed and the tests we have conducted.