ISSN 0253-2778

CN 34-1054/N

2014 Vol. 44, No. 10

Display Method:
Original Paper
A gait pattern planning algorithm based on linear coupled oscillator model for humanoid robots
WANG Liuqing, SHANG Weiwei
2014, 44(10): 795-803. doi: 10.3969/j.issn.0253-2778.2014.10.001
For the gait pattern planning algorithm of humanoid robots with the linear coupled oscillator model, some oscillator parameters were chosen by manual adjustment, which makes it difficult to get the parameters values that could ensure stable walking. To reduce the time of parameter selection and simplify the gait planning algorithm, a parameter selection algorithm based on linear coupled model for humanoid robots was proposed. By combining the walking speed, oscillation amplitude of the mass center during walking, and the trajectory curvature of ZMP, the optimization goal was established. The stable margin was defined and chosen as a constraint condition, and the nonlinear constraint method was used to obtain the optimization parameters which could ensure stable walking at faster speed. Simulation experiment demonstrated that the optimization parameters could ensure stable walking with rich stable margin. The proposed method was implemented on the open humanoid platform DARwIn-OP, and the results indicate that the optimized parameters in the linear coupled oscillator can achieve stable walking with fast gait pattern generation.
A digital control approach for unstable-first-order- plus-dead-time systems with P and PI controllers
LENG Tongtong, YAN Peng
2014, 44(10): 804-810. doi: 10.3969/j.issn.0253-2778.2014.10.002
Many dynamical systems can be modeled by unstable-first-order-plus-dead-time (UFOPDT) transfer functions. However, analysis and synthesis of UFOPDT systems are much more challenging due to the general difficulties of infinite dimensionality and the instability of the plant. Considering the control of such systems, explicit tuning formulae were derived for proportional (P) and proportional-integral (PI) controllers, based on the digitized open loop systems. Stability range was also discussed for the feedback systems with delays. Compared with existing results, the presented method significantly improved the accuracy and sufficiency, and simplified the tuning process. Numerical example about an isothermal chemical reactor control problem was given to illustrate this algorithm, and several relevant methods were also compared.
Markovian quantum subsystems and stabilization design
YAO Xiang, KUANG Sen
2014, 44(10): 811-817. doi: 10.3969/j.issn.0253-2778.2014.10.003
Quantum subsystems have been widely applied in the field of quantum information processing and control, so active control over quantum subsystems is of great importance. For Markovian open quantum dynamical model, some main theoretical results on invariant and attractive subsystems are summarized, and some important properties and their essential meanings are analyzed. Based on these, open-loop Hamiltonian design methods which make a subsystem become invariant and attractive were studied. In particular, the Hamiltonian design of invariance of a subsystem and attractiveness of an invariant subsystem achieved globally asymptotical stabilization of this subsystem. Finally, simulation experiments on a two-level open quantum system were conducted, and results show the validity of the proposed method.
Design and analysis of a universal model reference adaptive controller based on extended state observer
LI Jie, QI Xiaohui, ZHEN Hongtao, LIU Xinhai
2014, 44(10): 818-827. doi: 10.3969/j.issn.0253-2778.2014.10.004
A universal model reference adaptive control (MRAC) method based on extended state observer (ESO) was presented, which was used to directly estimate states and uncertainties and then compensate the uncertainties. Benefiting from the ESO, we only need to know the order of a system without distinction between linear and nonlinear system, time-varying and time-invariant system, the internal (parameter or structure) uncertainty and the external (disturbance) uncertainty etc. Furthermore, the reference model which meets desired performance index and has the same order with the system can be chosen arbitrarily largely independently of the system structure. Therefore, the controller design process is greatly simplified. After proposing the control architecture, the control law was designed and a strict stability analysis was given. The features mentioned above in addition to strong robustness, small control moment and high steady state accuracy are demonstrated by the simulation results.
Mobile location tracking based on NLOS identification in indoor environments
ZHANG Yanlong, CHEN Weidong, CHEN Chang
2014, 44(10): 828-834. doi: 10.3969/j.issn.0253-2778.2014.10.005
NLOS propagation is one of the key factors that affect tracking accuracy in indoor environments. An adaptive tracking algorithm was proposed to mitigate the NLOS error for indoor mobile localization. The correlation between adjacent NLOS errors in time was analyzed and exploited. A modified extended Kalman filter (MEKF) was presented which includes the NLOS error as part of the state variables. NLOS identification was achieved based on the state estimation of MEKF. MEKF and NLOS identification were combined to implement the adaptive tracking algorithm. Simulation results demonstrate that the proposed algorithm has better tracking accuracy and adaptability in indoor environments.
Constructing hash function from plaintext to Huff curves
YU Wei, WANG Kunpeng, LI Bao
2014, 44(10): 835-838. doi: 10.3969/j.issn.0253-2778.2014.10.006
A deterministic function with its time complexity being O(log3q) was constructed for the first time from a finite field Fq to Huff curves. Based on this function, construct a hash function from plaintext into Huff curves. The Hash function laid the foundation for identity-based protocols on elliptic curves and its experimental time cost on the elliptic curve of the Chinese elliptic curve standard SM2 is 5578 μs.
A detection algorithm based on matrix factorization for live mitochondria in fluorescent microscopic images
JIA Xiaomeng, NIE Yongming, DU Sidan
2014, 44(10): 839-843. doi: 10.3969/j.issn.0253-2778.2014.10.007
Detection of mitochondria in fluorescent microscopic images is one of the most important methods in studies concerning apoptosis and the nature of life phenomena in the area of biomedical image processing. Limited by fluorescence microscopy, fluorescent microscopic images contain two parts which are the shadow of cytoplasm and live mitochondria, and the signal-to-noise ratio (SNR) of live mitochondria time sequence images is low, which cannot meet the requirements of general particle algorithm. A new detection algorithm was proposed for live mitochondria in fluorescent microscopic images. To realize this method rapidly, augmented Lagrange multiplier algorithm was used. Mitochondria was be separated from the cytoplasm and accurately detected in fluorescent microscopic images. Therefore, the proposed algorithm provides an efficient and accurate tool to detect mitochondria in live cell.
A robust approximate algorithm for large-scale clustering of visual features
LI Darui, YANG Linjun, HUA Xiansheng, ZHANG Hongjiang
2014, 44(10): 844-852. doi: 10.3969/j.issn.0253-2778.2014.10.008
The large-scale clustering problem of visual features is crucial for image recognition and retrieval. The state-of-the-art algorithm, the approximate k-means, approximately guarantees the clustering performance by applying the high-precision approximate search. An improved algorithm was proposed, which requires no extra memory cost and nearly no extra time consumption. The robust approximate algorithm can better guarantee its convergence and clustering performance by utilizing more information in the iteration to update the partition, so that clustering loss is non-increasing and reduced rapidly. Theoretical proofs guarantee that the algorithm converges to the converged solution of the Lloyd algorithm, regard less of the precision values of approximate search. The experiment results show that the algorithm has about 10 times the speed of the approximate k-means algorithm. Besides, the clustering performance is also directly verified by comparing the images in the clustering results of global features.
An approach to evaluate the effectiveness of privacy protection in Android system
ZENG Shuke, ZHANG Yang, CHENG Liang, DENG Yi, FENG Dengguo
2014, 44(10): 853-861. doi: 10.3969/j.issn.0253-2778.2014.10.009
To protect private data in smart phones, Android enforces a permission-based security policy. PrivacyMiner, a tool for evaluating the effectiveness of privacy protection in Android, was designed. First, 22 categories of private data in smart phones were identified, which were then checked to see if Android could efficiently protect them from malware. PrivacyMiner was applied to 12 revisions of Android source code, and it was found that 7 categories of private data were not well protected, as Malware can read them and send them out without any permission. These vulnerabilities were verified on 6 Android devices with 6 revisions of Android, from 2.1 up to 4.4.2. Our findings were confirmed by the Android Security Team from Google.
Research on misjudgment of inter-AP handoff for on-board Wi-Fi applications
KANG Yuwen, ZOU Fumin, KANG Xingbin, ZHU Quan
2014, 44(10): 862-866. doi: 10.3969/j.issn.0253-2778.2014.10.010
At different scan intervals and moving speeds, simulation and experimental studies have been carried out for the misjudgment performance of inter-AP handoff for on-board Wi-Fi applications. The results show that, in a fast-moving environment, if the scan interval is reasonable, the misjudgment rate of inter-AP handoff is low, which can meet the requirements for handoff judgment of on-board Wi-Fi applications.
Asynchronous parallelism weighted A* algorithm for the finding shortest path on road networks
LENG Xuntai, SUN Guangzhong
2014, 44(10): 867-873. doi: 10.3969/j.issn.0253-2778.2014.10.011
Finding the shortest path is a classic problem with numerous applications. For road networks, it is desirable to find a sufficient short path within a limited period of time, and find a shorter path if there is more time. Since the traditional shortest path algorithms did not consider this constraint when designed, they can not meet the application requirement. To address this issue, an algorithm called APWA*, asynchronous parallelism weighted A*, was proposed, which can respond to users interrupt signal and return to the currently shortest path. Experiments on multiple maps show APWA* can meet the application requirement.
An efficient lower-bounding approach to point-to-point shortest path problem
ZHANG Zhong, LV Min, SUN Guangzhong, CHEN Guoliang
2014, 44(10): 874-880. doi: 10.3969/j.issn.0253-2778.2014.10.012
Querying for the shortest path from a source vertex to a sink vertex in real time is a fundamental problem in numerous applications. Several lower-bounding schemes have been proposed to solve the problem so far, such as A* search and ALT algorithm. But these schemes adopted loose valuations on distance so that there are great potentialities for improving the lower bounds. A novel two-stage goal directed lower-bounding approach, called ACT algorithm, was proposed, which combined A* search, centers and triangle inequality with no prior domain knowledge. Unlike previous schemes, the new algorithm can obtain excellent distance bounds by exploiting a large number of pre-computed data. The experimental results on real road networks show that ACT algorithm significantly outperforms all previous algorithms. In some instances, the vertices scanned by ACT algorithm are only about 25% more than those on optimal paths.