Yongmin Zhong

Random weighting estimation for fusion of multi-dimensional position data

This paper adopts the concept of random weighting estimation to multi-sensor data fusion. It presents a new random weighting estimation methodology for optimal fusion of multi-dimensional position data. A multi-sensor observation model is constructed for multi-dimensional position. Based on this observation model, a random weighting estimation algorithm is developed for estimation of position data from single sensors. Using the random weighting estimations from each single sensor, an optimization theory is established for optimal fusion of multi-sensor position data. Experimental results demonstrate that the proposed methodology can effectively fuse multi-sensor dimensional position data, and the fusion accuracy is much higher than that of the Kalman fusion method.

Random Weighting Method for Multisensor Data Fusion

This paper presents a new data fusion method by adopting random weighting estimation for optimal weighted fusion of multisensor observation data. This method adjusts in real time the weights of individual sensors according to variations in estimated sensor variances to obtain optimal weight distribution. Theories of random weighting estimation are established for optimal data fusion through optimal weighting distribution. Algorithms of random weighting estimation are developed to calculate sensor variances for determination of optimal random weighting factors. The fusion result in least mean square error is achieved directly from multisensor observation data, without requirement of any prior knowledge on unknown parameters. The mean square error estimated by the proposed method is not only smaller than from each individual sensor, but also smaller than by the mean of multisensor observation data.

An electromechanical based deformable model for soft tissue simulation

OBJECTIVEnSoft tissue deformation is of great importance to surgery simulation. Although a significant amount of research efforts have been dedicated to simulating the behaviours of soft tissues, modelling of soft tissue deformation is still a challenging problem. This paper presents a new deformable model for simulation of soft tissue deformation from the electromechanical viewpoint of soft tissues.nnnMETHODS AND MATERIALnSoft tissue deformation is formulated as a reaction-diffusion process coupled with a mechanical load. The mechanical load applied to a soft tissue to cause a deformation is incorporated into the reaction-diffusion system, and consequently distributed among mass points of the soft tissue. Reaction-diffusion of mechanical load and non-rigid mechanics of motion are combined to govern the simulation dynamics of soft tissue deformation.nnnRESULTSnAn improved reaction-diffusion model is developed to describe the distribution of the mechanical load in soft tissues. A three-layer artificial cellular neural network is constructed to solve the reaction-diffusion model for real-time simulation of soft tissue deformation. A gradient based method is established to derive internal forces from the distribution of the mechanical load. Integration with a haptic device has also been achieved to simulate soft tissue deformation with haptic feedback.nnnCONCLUSIONSnThe proposed methodology does not only predict the typical behaviours of living tissues, but it also accepts both local and large-range deformations. It also accommodates isotropic, anisotropic and inhomogeneous deformations by simple modification of diffusion coefficients.

Thermal–Mechanical-Based Soft Tissue Deformation for Surgery Simulation

Soft tissue deformation is of great importance to virtual reality-based surgery simulation. This paper presents a new methodology for the modeling of soft tissue deformation. This methodology converts soft tissue deformation into thermal–mechanical interaction according to the continuum mixture theory of soft tissues, and thus heat conduction of mechanical load and non-rigid mechanics of motion are combined to govern the dynamics of soft tissue deformation. The mechanical load applied to a soft tissue to cause a deformation is distributed among mass points of the soft tissue according to the principle of heat conduction. A thermal–mechanical model and associated model construction algorithms are developed to describe the distribution of the mechanical load in the tissue. A heat flux-based method is established for derivation of internal forces from the distribution of the mechanical load. Real-time interactive deformation of virtual human organs with force feedback has been achieved by the proposed methodology for surgery simulation. The proposed methodology not only accommodates isotropic, anisotropic and inhomogeneous materials by simply modifying thermal conductivity constants, but it also accepts local and large-range deformation.

Modelling a precision loadcell using neural networks for vision-based force measurement in cell micromanipulation

This paper presents a vision-based method to model a precision loadcell with artificial neural networks. The proposed model is used for measuring the applied force to a spherical biological cell during micromanipulation processes. The devised vision-based method is most useful where force feedback is required while integrating a force sensor into a cell micromanipulation setup is a challenging job. The proposed neural network model is used in conjunction with a methodology to track and characterize the cell deformation by extracting a geometric feature referred to as the `dimple angle directly from images of the cell micromanipulation process. The neural network is trained and used for the experimental data of zebrafish embryos micromanipulation. However, the proposed neural network is applicable for indentation of any other spherical elastic object. The results demonstrate the capability of the proposed method. The outcomes of this study could be useful for measuring force in biological cell microinjection processes such as injection of the mouse oocyte/embryo.

An optimal parameter estimation method for soft tissue characterization

This paper presents a gradient-free direct search estimation method by using genetic algorithm to model and predict the elastic stress response of ligament based on quasi-linear viscoelastic (QLV) theory. An improved genetic algorithm is developed to simultaneously fit the ramping and relaxation experimental data to the QLV constitutive equation for obtaining soft tissue parameters in a time-saving process. Experiments and comparison analysis with the existing methods for two exponential and polynomial QLV models are conducted, demonstrating that the proposed method can accurately estimate soft tissue parameters and satisfy the time-saving requirement of intra-operative soft tissue characterization.

An Improved Approach to Estimate Soft Tissue Parameters Using Genetic Algorithm for Minimally Invasive Measurement

This paper evaluates the ability of a gradient-free estimation method using genetic algorithm (GA) to model the elastic stress response of the anterior cruciate ligament (ACL) based on quasi-linear viscoelastic (QLV) theory. The improved GA simultaneously fits the ramping and relaxation experimental data to the QLV constitutive equation to obtain the soft tissue parameters. This approach is then compared with a previously evaluated method for two exponential and polynomial QLV models. The earlier approaches are mainly based on regression algorithms, which usually try to find a gradient-based solution with probability of poor convergence and variability of constants. Contrarily, this paper presents a gradient-free algorithm based on the improved timesaving GA. The results demonstrate that the ability of this algorithm to estimate the QLV parameters in timesaving process is functional to develop the optimal methodology for minimally invasive measurement during surgery.

Optimal Robot Path Planning with Cellular Neural Network

This paper presents a new methodology based on neural dynamics for optimal robot path planning by drawing an analogy between cellular neural network CNN and path planning of mobile robots. The target activity is treated as an energy source injected into the neural system and is propagated through the local connectivity of cells in the state space by neural dynamics. By formulating the local connectivity of cells as the local interaction of harmonic functions, an improved CNN model is established to propagate the target activity within the state space in the manner of physical heat conduction, which guarantees that the target and obstacles remain at the peak and the bottom of the activity landscape of the neural network. The proposed methodology cannot only generate real-time, smooth, optimal, and collision-free paths without any prior knowledge of the dynamic environment, but it can also easily respond to the real-time changes in dynamic environments. Further, the proposed methodology is parameter-independent and has an appropriate physical meaning.