浏览全部资源
扫码关注微信
收稿：2018-04-24，
修回：2018-6-13，
纸质出版：2018-12-16
移动端阅览
目的
2
基于物理模拟的人体运动生成方法由于能够合成符合自然规律的运动片段，可实时响应环境的变化，且生成的物理运动不是机械性的重复，因此是近年来计算机动画和虚拟现实领域中最活跃的研究方向之一。然而人体物理模型具有高维、非线性及关节间强耦合性等特点，求解人体物理运动十分困难。反馈控制器常用于人体物理运动控制，求解时通常需要对多个目标函数加权求和，然而权重的设置需多次试验，烦杂耗时。针对运动控制器求解困难的问题，本文提出了一种面向反馈运动控制器的多目标求解方法。
方法
2
首先，对运动数据进行预处理并提取关键帧求解初始控制器，并设计一种改进的反馈控制机制；在此基础上，种群父代个体变异产生子代，采用禁选区域预筛选策略去除不满足约束的个体，并通过重采样获取新解；然后，通过物理仿真获得多目标适应度值，采用区域密度多层取优选取分布均匀的优秀个体作为下一代父代，并通过基于剪枝的多阶段物理求解算法决定是否进入下一阶段优化；经过多次迭代后获得物理控制器，从而生成具有反馈的人体物理运动。
结果
2
针对提出的方法，本文针对多个测试函数和物理运动分别进行实验：在测试函数实验中，本文分别采用经典的测试函数进行实验对比，在相同的迭代次数下，相比之前算法，本文算法中满足约束的优秀个体命中率更高，反转世距离更小，且最优解集的分布更加均匀；物理运动生成实验中，分别针对走路、跑步和翻滚等运动进行物理运动生成，与之前算法进行对比，本文算法可以更早地完成收敛，同时目标函数值更小，表明生成的运动效果更好。
结论
2
本文提出的进化求解方法可以生成不同运动的控制器，该控制器不仅可以生成物理运动，而且还具备外力干扰下保持平衡的能力，解决了运动控制器求解中多目标权重设置困难、优化时间长的问题；除此之外，本文算法还对具有约束的多目标问题具有较好的求解效果。
Objective
2
A physical-based animation synthesis can generate a human physical motion
which satisfies physical laws. Human physical motion is generated by responding to the environment in real-time and is not mechanically repetitive. Therefore
the human physical motion has been an interesting topic in the computer graphics and virtual reality fields in recent years. However
the human physical motion is difficult to generate given the high dimensionality
nonlinearity
and strong coupling of joints in human physical model. In addition
a feedback controller is frequently used to control human physical motion
especially in the diverse environment or under external forces. Multiple objective functions are typically designed during the process of solving the feedback controllers. Researchers apply optimization methods to solve the feedback controller. Multiple objective functions are frequently converted to a single objective function by utilizing the weighted sum method. However
inaccurate weights can easily result in failure of convergence considering the local traps. Therefore
the setting of weights is crucial to the direction of optimization
convergence time
and result. Thus
experiments become difficult
and these systems require dedicated technical developers. Owing to these problems
a multi-objective solving method is proposed for a feedback motion controller.
Method
2
We preprocess motion data and extract key frames to construct the initial controller. Furthermore
an improved feedback control mechanism is designed to reduce the difficulty of the constraint-solving problem. The parents of the population generate children after variation. However
many failure individuals do not satisfy the constraints in the children population. We utilize a forbidden region pre-filtering strategy to solve the abovementioned problem. This strategy adopts the support vector machine (SVM) with a radial basis kernel function (RBF) to remove these failure individuals. We replenish new children by re-sampling to supplement these removed individuals. We set several objective functions
including root
pose
and energy cost functions to measure the quality of every individual. Then
every individual controller is inputted to the physical simulation system
and multi-objective fitness values are obtained after simulations are completed. The regional density multi-layer optimization algorithm is adopted to select the excellent individuals
which show a uniform distribution to construct the next generation of parent individuals. Simultaneously
the SVM with kernel RBF is updated. We prune several individual controllers
which merely generate short physical motion
to obtain stable controllers
which can generate an extensive physical motion. Then
we decide whether to proceed to the next stage of optimization through multi-stage physical solving algorithm based on pruning. The optimized physical controller is obtained after many iterations. Finally
the optimal controller is applied to generate a human physical motion with feedback.
Result
2
The experiments are conducted on multiple test functions and physical motions on the basis of the proposed method. In the test function experiment
the classical test functions
including
$${\rm{LZ}}1$$
$${\rm{ZDT}}2$$
and
$${\rm{DTLZ4}}$$
functions
are used for experimental comparison. We also set several user-defined constraints to analyze the anti-constraint capability of a method. The proposed method achieves higher hit ratio of excellent individuals than the existing methods
thereby satisfying user-defined constraints in the children population under the same number of iterations. The final optimal solution set achieves a small inverted generational distance (IGD) and shows a uniform distribution. In the physical motion experiment
a human physical motion is generated for walking
running
and rolling. Several constraints are appointed; for example
the human physical model is forbidden from falling down and ricocheting off while simulating locomotion. The proposed method achieves a high convergence speed and obtains small objective function values than the existing methods. The results indicate that our generated physical motion has better performance than the other techniques.
Conclusion
2
The proposed method can generate various physical human motion controllers
which can not only generate physical motion but also maintain the balance of a physical model under external forces. The proposed evolutionary method can solve the problem of setting multi-objective weights and conducting extensive optimization. In addition
the proposed method has a favorable performance in terms of other multi-objective problems with constraints.
Xia S H, Wei Y, Wang Z Q. A survey of physics-based human motion simulation[J]. Journal of Computer Resarch and Development. 2010, 47(8):1354-1361.
夏时洪, 魏毅, 王兆其.人体运动仿真综述[J].计算机研究与发展, 2010, 47(8):1354-1361.
Zhao J J, Wei Y, Xia S H, et al. Survey of physics-based character animation[J]. Journal of Computer Resarch and Development. 2015, 52(12):2866-2878.
赵建军, 魏毅, 夏时洪, 等.基于物理的角色动画合成方法综述[J].计算机研究与发展, 2015, 52(12):2866-2878. [DOI:10.7544/issn1000-1239.2015.20140634]
Xia S H, Gao L, Lai Y K, et al. A survey on human performance capture and animation[J]. Journal of Computer Science and Technology, 2017, 32(3):536-554.[DOI:10.1007/s11390-017-1742-y]
Al Borno M, De Lasa M, Hertzmann A. Trajectory optimization for full-body movements with complex contacts[J]. IEEE Transactions on Visualization and Computer Graphics, 2013, 19(8):1405-1414.[DOI:10.1109/TVCG.2012.325]
Liu L B, Yin K K, Van De Panne M, et al. Sampling-based contact-rich motion control[J]. ACM Transactions on Graphics, 2010, 29(4):#128.[DOI:10.1145/1778765.1778865]
Ding K, Liu L B, Van De Panne M, et al. Learning reduced-order feedback policies for motion skills[C ] //Proceedings of the 14th ACM SIGGRAPH/Eurographics Symposium on Computer Animation. Los Angeles, California: ACM, 2015.[ DOI: 10.1145/2786784.2786802 http://dx.doi.org/10.1145/2786784.2786802 ]
Yin K K, Loken K, Van De Panne M. SIMBICON:simple biped locomotion control[J]. ACM Transactions on Graphics, 2007, 26(3):#105.[DOI:10.1145/1276377.1276509]
Lee Y, Kim S, Lee J. Data-driven biped control[J]. ACM Transactions on Graphics, 2010, 29(4):#129.[DOI:10.1145/1778765.1781155]
Mordatch I, De Lasa M, Hertzmann A. Robust physics-based locomotion using low-dimensional planning[J]. ACM Transactions on Graphics, 2010, 29(4):#71.[DOI:10.1145/1778765.1778808]
Zhao J J, Wei Y, Xia S H, et al. Locomotion synthesis for humanoid character based on a double inverted pendulum model[J]. Chinese Journal of Computers, 2014, 37(10):2187-2195.
赵建军, 魏毅, 夏时洪, 等.基于二阶倒立摆的人体运动合成[J].计算机学报, 2014, 37(10):2187-2195. [DOI:10.3724/SP.J.1016.2014.02187]
Liu L B, Yin K K, Guo B N. Improving sampling-based motion control[J]. Computer Graphics Forum, 2015, 34(2):415-423.[DOI:10.1111/cgf.12571]
Lv X L, Chai J X, Xia S H. Data-driven inverse dynamics for human motion[J]. ACM Transactions on Graphics, 2016, 35(6):#163.[DOI:10.1145/2980179.2982440]
Da Silva M, Abe Y, Popovic' J. Interactive simulation of stylized human locomotion[J].ACM Transactions on Graphics, 2008, 27(3):#82.[DOI:10.1145/1360612.1360681]
Liu L B, Yin K K, Van De Panne M, et al. Terrain runner:control, parameterization, composition, and planning for highly dynamic motions[J]. ACM Transactions on Graphics, 2012, 31(6):#154.[DOI:10.1145/2366145.2366173]
Lu M, Zhang Y K, Liu X P. Multi-skeleton-oriented and multi-style-oriented locomotion controller[J]. Journal of Image and Graphics, 2016, 21(11):1539-1550.
鲁明, 张迎凯, 刘晓平.面向多骨骼及多风格的行走运动控制器[J].中国图象图形学报, 2016, 21(11):1539-1550. [DOI:10.11834/jig.20161114]
Agrawal S, Van De Panne M. Pareto optimal control for natural and supernatural motions[C ] //Proceedings of Motion on Games. Dublin, Ireland: ACM, 2013: 7-16.[ DOI: 10.1145/2522628.2522902 http://dx.doi.org/10.1145/2522628.2522902 ]
Ha S, Liu C K. Iterative training of dynamic skills inspired by human coaching techniques[J]. ACM Transactions on Graphics, 2014, 34(1):#1.[DOI:10.1145/2682626]
Firmin M, Van De Panne M. Controller design for multi-skilled bipedal characters[J]. Computer Graphics Forum, 2015, 34(8):50-63.[DOI:10.1111/cgf.12607]
Agrawal S, Shen S, Van De Panne M. Diverse motion variations for physics-based character animation[C ] //The 12th ACM SIGGRAPH/Eurographics Symposium on Computer Animation. Anaheim, California: ACM, 2013: 37-44.[ DOI: 10.1145/2485895.2485907 http://dx.doi.org/10.1145/2485895.2485907 ]
Lee Y, Lee K, Kwon S S, et al. Push-recovery stability of biped locomotion[J]. ACM Transactions on Graphics, 2015, 34(6):#180.[DOI:10.1145/2816795.2818124]
Li Z Y, H T, Chen S M, et al. Overview of constrained optimization evolutionary algorithms[J]. Journal of Software, 2017, 28(6):1529-1546.
李智勇, 黄滔, 陈少淼, 等.约束优化进化算法综述[J].软件学报, 2017, 28(6):1529-1546. [DOI:10.13328/j.cnki.jos.005259]
Igel C, Hansen N, Roth S. Covariance matrix adaptation for multi-objective optimization[J]. Evolutionary Computation, 2007, 15(1):1-28.[DOI:10.1162/evco.2007.15.1.1]
Bringmann K, Friedrich T. Approximating the volume of unions and intersections of high-dimensional geometric objects[C ] //The 19th International Symposium on Algorithms and Computation. Gold Coast, Australia: Springer-Verlag, 2008: 436-447.[ DOI: 10.1007/978-3-540-92182-0_40 http://dx.doi.org/10.1007/978-3-540-92182-0_40 ]
Abe Y, Da Silva M, Popovic' J. Multiobjective control with frictional contacts[C]//The 2007 ACM SIGGRAPH/Eurographics Symposium on Computer animation. San Diego: ACM, 2007: 249-258.
Li H, Zhang Q F. Multiobjective optimization problems with complicated pareto sets, MOEA/D and NSGA-Ⅱ[J]. IEEE Transactions on Evolutionary Computation, 2009, 13(2):284-302.[DOI:10.1109/TEVC.2008.925798]
Zitzler E, Deb K, Thiele L. Comparison of multiobjective evolutionary algorithms:empirical results[J]. Evolutionary Computation, 2000, 8(2):173-195.[DOI:10.1162/106365600568202]
Deb K, Thiele L, Laumanns M, et al. Scalable multi-objective optimization test problems[C ] //Proceedings of the 2002 Congress on Evolutionary Computation. Honolulu, HI, USA: IEEE, 2002: 825-830.[ DOI: 10.1109/CEC.2002.1007032 http://dx.doi.org/10.1109/CEC.2002.1007032 ]
Wang H, Yen G G, Zhang X, Multiobjective particle swarm optimization based on pareto entropy[J]. Journal of Software, 2014, 25(5):1025-1050.
旺胡, Yen G G, 张鑫.基于Pareto熵的多目标粒子群优化算法[J].软件学报, 2014, 25(5):1025-1050. [DOI:10.13328/j.cnki.jos.004496]
相关作者
相关机构
京公网安备11010802024621
