This paper deals with finding an n-dimensional solution x to a syst.m of quadratic equations yi = ||2, 1 ≤ i ≤ m, which in general is known to be NP-hard. We put forth a novel procedure, that starts with a weighted ...
ISBN:
(纸本)9781510860964
This paper deals with finding an n-dimensional solution x to a syst.m of quadratic equations yi = ||2, 1 ≤ i ≤ m, which in general is known to be NP-hard. We put forth a novel procedure, that starts with a weighted maximal correlation initialization obtainable with a few power iterations, followed by successive refinements based on iteratively reweighted gradient-type iterations. The novel techniques distinguish themselves from prior works by the inclusion of a fresh (re)weighting regularization. For certain random measurement models, the proposed procedure returns the true solution x with high probability in time proportional to reading the data {(ai; yi)]1≤i≤m, provided that the number m of equations is some constant c > 0 times the number n of unknowns, that is, m ≤ cn. Empirically, the upshots of this contr.bution are: i) perfect signal recovery in the high-dimensional regime given only an information-theoretic limit number of equations; and, ii) (near-)optimal statistical accuracy in the presence of additive noise. Extensive numerical tests using both synthetic data and real images corroborate its improved signal recovery performance and computational efficiency relative to state-of-the-art approaches.
An adaptive switch gain time-varying sliding mode contr.ller is designed for the low speed servo syst.m in a contr.l moment gyroscope (CMG). By introducing the integral of speed error in the sliding mode surface, the ...
详细信息
ISBN:
(纸本)9781479970995
An adaptive switch gain time-varying sliding mode contr.ller is designed for the low speed servo syst.m in a contr.l moment gyroscope (CMG). By introducing the integral of speed error in the sliding mode surface, the error is suppressed during the whole operation. A time-varying sliding mode contr.ller based on error reference is proposed to solve the problem that normal time-varying sliding mode contr.l is only related to time and can only eliminate the reaching phase at initial time. The proposed contr.ller makes the sliding mode surface keep refreshing all the time. The global robustness is maintained. Then, an adaptive switch gain sliding mode contr.ller is proposed to deal with the syst.m vibration. This contr.ller has large gain for large error and small gain for small error. Simulation and experiment were carried out for the proposed CMG low speed servo syst.m contr.ller. The results showed that the proposed contr.ller have better performance in terms of smooth speed and robustness comparing with conventional PID contr.ller.
In this paper, we focus on the need for autonomous vehicles' path planning and decision making in the urban environment. In our modified method, traffic lanes, road edge and geographic information syst.m(GIS) are ...
详细信息
ISBN:
(纸本)9781467365970
In this paper, we focus on the need for autonomous vehicles' path planning and decision making in the urban environment. In our modified method, traffic lanes, road edge and geographic information syst.m(GIS) are integrated to generate three structured road models, which can be used to reduce car's lateral positioning error and guarantee car's driving stably without a precise pre-computed map for localization. Based on these models, car's position, attitude and speed are taken into account to find an optimal guided path and a suitable behavior strategy to keep our car obeying traffic rules. Thus, in the whole architecture, the instability of some modules' detection results and low positioning accuracy of GPS receiver is resolved. Applied to our autonomous vehicle - IN 2 BOT, which has taken part in "Chinese intell.gent Vehicle Challenge" for many times, this method is proved more effective than our previous method after experiments under various conditions.
In this paper, we introduce a collaborative target tracking method which couples distributed estimation and motion contr.l of mobile sensor networks. To estimate the target state, we develop a distributed federated fi...
详细信息
ISBN:
(纸本)9781479905607
In this paper, we introduce a collaborative target tracking method which couples distributed estimation and motion contr.l of mobile sensor networks. To estimate the target state, we develop a distributed federated filter (DFF) approach where only local communication between sensors is required. This DFF algorithm has superiority in tolerance of the target syst.m noise and can acquire accurate estimation even if the model of the target is not accurate. We value information-distribution and information-fusion weights in the DFF according to the measurement situations of sensors. Then a flocking-based motion contr.l method is proposed to track maneuvering targets. During the process of tracking, velocity matching and collision avoidance among mobile sensors are guaranteed. Finally, this coupled estimation and contr.l algorithm is successfully simulated to track a linear-dynamics target with non-zero acceleration. Simulation results verify the performance and superiority of our method.
暂无评论