{"title":"Unsupervised inverse reinforcement learning with noisy data","authors":"A. Surana","doi":"10.1109/CDC.2014.7040160","DOIUrl":"https://doi.org/10.1109/CDC.2014.7040160","url":null,"abstract":"In this paper we propose an approach for unsupervised Inverse Reinforcement Learning (IRL) with noisy data using a hidden variable Markov Decision Processes (hMDP) representation. hMDP accounts for observation uncertainty by using a hidden state variable. We develop a nonparametric Bayesian IRL technique for hMDP based on Dirichlet Processes mixture model. We provide an efficient Markov Chain Monte Carlo based sampling algorithm whereby one can automatically cluster noisy data into different behaviors, and estimate the underlying reward parameters per cluster. We demonstrate our approach for unsupervised learning, and prediction and classification of agent behaviors in a simulated surveillance scenario.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129549983","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
G. Kurz, Igor Gilitschenski, Maxim Dolgov, U. Hanebeck
{"title":"Bivariate angular estimation under consideration of dependencies using directional statistics","authors":"G. Kurz, Igor Gilitschenski, Maxim Dolgov, U. Hanebeck","doi":"10.1109/CDC.2014.7039789","DOIUrl":"https://doi.org/10.1109/CDC.2014.7039789","url":null,"abstract":"Estimation of angular quantities is a widespread issue, but standard approaches neglect the true topology of the problem and approximate directional with linear uncertainties. In recent years, novel approaches based on directional statistics have been proposed. However, these approaches have been unable to consider arbitrary circular correlations between multiple angles so far. For this reason, we propose a novel recursive filtering scheme that is capable of estimating multiple angles even if they are dependent, while correctly describing their circular correlation. The proposed approach is based on toroidal probability distributions and a circular correlation coefficient. We demonstrate the superiority to a standard approach based on the Kalman filter in simulations.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128547468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Chao Zhai, Francesco Alderisio, K. Tsaneva-Atanasova, M. Bernardo
{"title":"Adaptive tracking control of a virtual player in the mirror game","authors":"Chao Zhai, Francesco Alderisio, K. Tsaneva-Atanasova, M. Bernardo","doi":"10.1109/CDC.2014.7040490","DOIUrl":"https://doi.org/10.1109/CDC.2014.7040490","url":null,"abstract":"The coordination of interpersonal rhythmic movements is of great significance due to its potential relevance to human motor rehabilitation. In this paper we consider the problem of designing a controller able to drive a virtual player (VP) capable of imitating and following a human player in the mirror game [1]. The classic nonlinear Haken-Kelso-Bunz (HKB) model is adopted to describe the social motor coordination between two players. An adaptive control algorithm is developed and implemented on the HKB model to drive the VP. It is proven that the position error between the VP driven by our control algorithm and the human player is upper bounded during the game. Finally, experiments are conducted on a prototype set-up in order to evaluate the performance of the proposed control algorithm and compare it with other existing algorithms.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129154507","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Dynamics in network games with local coordination and global congestion effects","authors":"Gianluca Brero, G. Como, F. Fagnani","doi":"10.1109/CDC.2014.7039708","DOIUrl":"https://doi.org/10.1109/CDC.2014.7039708","url":null,"abstract":"Several strategic interactions over social networks display both negative and positive externalities at the same time. E.g., participation to a social media website with limited resources is more appealing the more of your friends participate, while a large total number of participants may slow down the website (because of congestion effects) thus making it less appealing. Similarly, while there are often incentives to choose the same telephone company as the friends and relatives with whom you interact the most frequently, concentration of the market share in the hands of a single firm typically leads to higher costs because of the lack of competition. In this work, we study evolutionary dynamics in network games where the payoff of each player is influenced both by the actions of her neighbors in the network, and by the aggregate of the actions of all the players in the network. In particular, we consider cases where the payoff increases in the number of neighbors who choose the same action (local coordination effect) and decreases in the total number of players choosing the same action (global congestion effect). We study noisy best-response dynamics in networks which are the union of two complete graphs, and prove that the asymptotic behavior of the invariant probability distribution is characterized by two phase transitions with respect to a parameter measuring the relative strength of the local coordination with respect to the global congestion effects. Extensions to random networks with strong community structure are studied through simulations.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129333358","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Generating function approach to linear quadratic optimal control problem with constraints on the state","authors":"Dijian Chen, K. Fujimoto, Tatsuya Suzuki","doi":"10.1109/CDC.2014.7040434","DOIUrl":"https://doi.org/10.1109/CDC.2014.7040434","url":null,"abstract":"This paper extends the generating function approach to the state constrained linear quadratic optimal control problem by utilizing penalties. The original constrained problem is approximated by an unconstrained problem with a novel penalty function by which the latter problem can be solved via generating function method based on Taylor series approximation. In view of this, the optimal solution can be given as the state feedback control with explicit boundary conditions and pre-computed coefficients. This structure is useful in the repetitive on-line computation for optimal solutions satisfying a number of different boundary conditions. Examples demonstrate the effectiveness of the developed method.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124547261","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Junkai Lu, Mo Chen, Y. Chang, M. Tomizuka, J. Carmena, C. Tomlin
{"title":"Design of a neural decoder by sensory prediction and error correction","authors":"Junkai Lu, Mo Chen, Y. Chang, M. Tomizuka, J. Carmena, C. Tomlin","doi":"10.1109/CDC.2014.7040489","DOIUrl":"https://doi.org/10.1109/CDC.2014.7040489","url":null,"abstract":"Brain-machine interfaces (BMI) hold great potential to improve the quality of life of many patients with disabilities. The neural decoder, which expresses the mapping between the neural signals and the subject's motion, plays an important role in BMI systems. Conventional neural decoders are generally in the form of a kinematic Kalman filter which does not possess an explicit mechanism to deal with the unavoidable mismatch between the biological system and the model of the system used by the decoder. This paper presents a novel design of a neural decoder that uses a one-step model predictive controller to generate a control signal that compensates for the inherent model mismatch. The effectiveness of the proposed decoding algorithm compares favorably to the state-of-the-art Kalman filter in numerical simulations with different degrees of model mismatch.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129832649","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Extremum-seeking control of distributed systems using consensus estimation","authors":"S. Dougherty, M. Guay","doi":"10.1109/CDC.2014.7039924","DOIUrl":"https://doi.org/10.1109/CDC.2014.7039924","url":null,"abstract":"A method for solving constrained optimization problems in real-time using a distributed control system is presented in this work. The actions of a controller is based on measurements of a part of the total cost function. The algorithm is classified as an extremum-seeking control approach because the design requires no knowledge of the cost function. The controllers coordinate movements by solving a consensus estimation problem in parallel with the optimization. A proof for the algorithm's convergence is presented. Two simulation examples are used to study the behaviour of the proposed system.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130376877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Frame induced vibration estimation and attenuation scheme on a multirotor helicopter","authors":"Emil Fresk, G. Nikolakopoulos","doi":"10.1109/CDC.2014.7040281","DOIUrl":"https://doi.org/10.1109/CDC.2014.7040281","url":null,"abstract":"The aim of this article is to establish an induced frame vibration and attenuation scheme, specifically targeting the area of multi-rotor Unmanned Aerial Vehicles (UAVs), such as quadrotors. These types of unmanned small scale helicopters are characterized by small and light frames, which are vulnerable to vibrations induced by the operation of the motors or external environmental factors. The existence of such vibrations effecting the frame can significantly deteriorate the performance of the overall closed system and even drive it to instabilities. In this article spectral estimation schemes based on: a) Autoregressive (AR) modeling and b) Multiple Signal Classification (MUSIC) are being established and evaluated towards their ability to detect the induced vibration frequencies on the UAV, while an extended discussion is being presented on selecting the correct number of the identified induced vibrating frequencies. In a sequential stage, a vibration attenuation approach based on notch filtering is being presented, being able to correctly attenuate the induced vibrating frequencies in the measurements. The efficiency of the overall suggested scheme is being evaluated by experimental results that indicate the significant improvement in the measurements achieved by the direct application of the proposed scheme.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129092126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
N. Amelina, O. Granichin, O. Granichina, Yuming Jiang
{"title":"Differentiated consensuses in decentralized load balancing problem with randomized topology, noise, and delays","authors":"N. Amelina, O. Granichin, O. Granichina, Yuming Jiang","doi":"10.1109/CDC.2014.7040484","DOIUrl":"https://doi.org/10.1109/CDC.2014.7040484","url":null,"abstract":"In this paper, a new consensus problem, termed differentiated consensuses, is studied. This consensus problem is that, in a system with multiple classes, consensus is targeted for each class, which may be different among classes. Specifically, we investigate differentiated consensuses in a distributed stochastic network system of nodes (or agents), where tasks, classified with different priorities, are serviced. The network system is assumed to have switched topology, noise and delay in measurement, and cost on the topology. The goal is to reach/maintain balanced (equal) load, i.e. consensus, across the network and at the same time to meet the topology cost constraint, both for every priority class. A control protocol is proposed. We prove that the proposed control protocol is able to meet the topology cost constraint and achieve approximate consensus for each of the priority classes in the network.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123851023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jose Louis Avila Alonso, C. Bonnet, E. Fridman, F. Mazenc, J. Clairambault
{"title":"Stability analysis of PDEs modelling cell dynamics in Acute Myeloid Leukemia","authors":"Jose Louis Avila Alonso, C. Bonnet, E. Fridman, F. Mazenc, J. Clairambault","doi":"10.1109/CDC.2014.7039860","DOIUrl":"https://doi.org/10.1109/CDC.2014.7039860","url":null,"abstract":"In this paper we perform a stability analysis of two systems of partial differential equations (PDEs) modelling cell dynamics in Acute Myeloid Leukemia. By using a Lyapunov approach, for an equilibrium point of interest, we obtain stability bounds depending on the parameters of the systems. First, we derive sufficient conditions for boundedness of solutions. Then, asymptotic stability conditions are obtained. The results are illustrated with numerical examples and simulations.","PeriodicalId":202708,"journal":{"name":"53rd IEEE Conference on Decision and Control","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123192867","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}