In this paper, we study optimal LQG tracking control problems for linear continuous-time hybrid systems. We adopt an approach to prepare some candidates of mode distributions and minimize value functions at each time for the systems and performance indices. By the approach, system states and mode distributions are simultaneously optimized so that the performance indices are minimized for available information of reference signals. We consider optimization problems for averaged systems and averaged performance indices throughout mode distributions. Finally we give numerical examples and verify that we can obtain both optimal tracking performance and optimal candidate of mode distributions by the algorithm presented in this paper.
In wireless LANs (Local Area Networks), each station adapts its transmission rate according to the channel condition. Thus stations with different transmission rates coexist in a LAN. In such a LAN, the standard access method DCF (Distributed Coordination Function) causes the problem called performance anomaly, all the stations have the same throughput in spite of their transmission rates since the DCF provides equal access privileges among stations. In order to solve the problem, this paper proposes an access method which achieve transmission time fairness. Under the proposed method, the access point calculates the backoff periods for all stations. That could eliminate packet collisions and redundant idle periods. Via simulation, we show effectiveness of our method in comparison with the existing ones.
In recent years autonomous control of automobiles has been focused on. Path-following control is an important technologies of autonomous control of automobiles. In path-following control, it is necessary to calculate an appropriate steering input and velocity input according to the shape of the target path. In previous research, the steering control and velocity control method have been achieved by formulating and solving the optimization problem using variation principle. Since the problem solvable by the two-point boundary value problem is limited, the format of the evaluation function is limited and it is difficult to explicitly consider constraints in some cases. In addition, the real-time performance of the optimization calculation is also a problem considering the implementation in the real plant. In this paper, we propose a method for accurate follow-up to the target path and appropriate velocity control according to the shape of the target path in real time by designing the velocity function as a function of the target path information and state of the plant.