An approach of traffic signal control based on NLRSQP algorithm

This paper presents a linear program model with linear complementarity constraints (LPLCC) to solve traffic signal optimization problem. The objective function of the model is to obtain the minimization of total queue length with weight factors at the end of each cycle. Then, a combination algorithm based on the nonlinear least regression and sequence quadratic program (NLRSQP) is proposed, by which the local optimal solution can be obtained. Furthermore, four numerical experiments are proposed to study how to set the initial solution of the algorithm that can get a better local optimal solution more quickly. In particular, the results of numerical experiments show that: The model is effective for different arrival rates and weight factors; and the lower bound of the initial solution is, the better optimal solution can be obtained.

Dynamic Simulation of Autonomous Boats for Cooperative Skimming and Cleanup

We consider the use of autonomous boats for oil skimming and clean up of surface debris by operating two boats with a boom cooperatively. A boom, or a cable as considered in this paper, attached to two robots at each end can be used to efficiently manipulate multiple objects on the surface of the water. In our previous work, Bhattacharya, et al. (ICRA 2011), we showed the feasibility of this concept with an experimental testbed using two autonomous boats and a towed cable. This work showed that an accurate dynamic simulation of the system is indispensable in analysis and development of efficient control schemes. For the purpose of manipulating objects in such a way, not only do we need to model the drag forces, but we also need to model the interaction between the cable and the objects. In this paper we model the boom or the cable as a chain with a discrete number of rigid links connected by passive revolute joints and model the interaction of the cable with the water (drag) as well as the contacts with objects on the surfaces. We derive the equations governing the cable and object dynamics and model the contact interactions as linear complementarity constraints. The boats are driven by simple controllers that only require knowledge of positions and velocities at the both ends of the cable. Several examples are used to illustrate the performance of the numerical simulation.