与传统的一阶马尔科夫随机场(Markov random field,MRF)相比,高阶马尔科夫随机场能够表达更加复杂的定性和统计性先验信息,在模型的表达能力上具有更大的优势.但高阶马尔科夫随机场对应的能量函数优化问题更为复杂.同时其模型参数数目...与传统的一阶马尔科夫随机场(Markov random field,MRF)相比,高阶马尔科夫随机场能够表达更加复杂的定性和统计性先验信息,在模型的表达能力上具有更大的优势.但高阶马尔科夫随机场对应的能量函数优化问题更为复杂.同时其模型参数数目的爆炸式增长使得选择合适的模型参数也成为了一个非常困难的问题.近年来,学术界在高阶马尔科夫随机场的能量模型的建模、优化和参数学习三个方面进行了深入的探索,取得了很多有意义的成果.本文首先从这三个方面总结和介绍了目前在高阶马尔科夫随机场研究上取得的主要成果,然后介绍了高阶马尔科夫随机场在图像理解和三维场景理解中的应用现状.展开更多
A prescribed performance neural controller to guarantee tracking quality is addressed for the near space kinetic kill vehicle (NSKKV) to meet the state constraints caused by side window detection. Different from the t...A prescribed performance neural controller to guarantee tracking quality is addressed for the near space kinetic kill vehicle (NSKKV) to meet the state constraints caused by side window detection. Different from the traditional prescribed performance control in which the shape of the performance function is constant, this paper exploits new performance functions which can change the shape of their function according to different symbols of initial errors and can ensure the error convergence with a small overshoot. The neural backstepping control and the minimal learning parameters (MLP) technology are employed for exploring a prescribed performance controller (PPC) that provides robust tracking attitude reference trajectories. The highlight is that the transient performance of tracking errors is satisfactory and the computational load of neural approximation is low. The pseudo rate (PSR) modulator is used to shape the continuous control command to pulse or on-off signals to meet the requirements of the thruster. Numerical simulations show that the proposed method can achieve state constraints, pseudo-linear operation and high accuracy.展开更多
文摘与传统的一阶马尔科夫随机场(Markov random field,MRF)相比,高阶马尔科夫随机场能够表达更加复杂的定性和统计性先验信息,在模型的表达能力上具有更大的优势.但高阶马尔科夫随机场对应的能量函数优化问题更为复杂.同时其模型参数数目的爆炸式增长使得选择合适的模型参数也成为了一个非常困难的问题.近年来,学术界在高阶马尔科夫随机场的能量模型的建模、优化和参数学习三个方面进行了深入的探索,取得了很多有意义的成果.本文首先从这三个方面总结和介绍了目前在高阶马尔科夫随机场研究上取得的主要成果,然后介绍了高阶马尔科夫随机场在图像理解和三维场景理解中的应用现状.
基金supported by the National Natural Science Foundation of China(61773398 61703421)
文摘A prescribed performance neural controller to guarantee tracking quality is addressed for the near space kinetic kill vehicle (NSKKV) to meet the state constraints caused by side window detection. Different from the traditional prescribed performance control in which the shape of the performance function is constant, this paper exploits new performance functions which can change the shape of their function according to different symbols of initial errors and can ensure the error convergence with a small overshoot. The neural backstepping control and the minimal learning parameters (MLP) technology are employed for exploring a prescribed performance controller (PPC) that provides robust tracking attitude reference trajectories. The highlight is that the transient performance of tracking errors is satisfactory and the computational load of neural approximation is low. The pseudo rate (PSR) modulator is used to shape the continuous control command to pulse or on-off signals to meet the requirements of the thruster. Numerical simulations show that the proposed method can achieve state constraints, pseudo-linear operation and high accuracy.
