Fan Liang

Heart Motion Uncertainty Compensation Prediction Method for Robot Assisted Beating Heart Surgery --- Master---slave Kalman Filters Approach

Robot Assisted Coronary Artery Bypass Graft (CABG) allows the heart keep beating in the surgery by actively eliminating the relative motion between point of interest (POI) on the heart surface and surgical tool. The inherited nonlinear and diverse nature of beating heart motion gives a huge obstacle for the robot to meet the demanding tracking control requirements. In this paper, we novelty propose a Master–slave Kalman Filter based on beating heart motion Nonlinear Adaptive Prediction (NAP) algorithm. In the study, we describe the beating heart motion as the combination of nonlinearity relating mathematics part and uncertainty relating non-mathematics part. Specifically, first, we model the nonlinearity of the heart motion via quadratic modulated sinusoids and estimate it by a Master Kalman Filter. Second, we involve the uncertainty heart motion by adaptively change the covariance of the process noise through the slave Kalman Filter. We conduct comparative experiments to evaluate the proposed approach with four distinguished datasets. The results indicate that the new approach reduces prediction errors by at least 30 μm. Moreover, the new approach performs well in robustness test, in which two kinds of arrhythmia datasets from MIT-BIH arrhythmia database are assessed.

Design of Intelligent Measure and Control Software of Plant Growth Cabinet Based on Android System

Plant growth cabinet, as a kind of green, environmental, efficient, energy-efficient, closed plant growth environment, can provide the temperature, humidity and others that are required in the process of plant growth and development. We design an intelligent measure and control software based on Android system to achieve information collection and display of the environmental data in the cabinet and the plants’ morphological data. It also could remotely monitor the plant growth process in real time and do the intelligent regulation of the control node. In order to collect and send environmental data and video in real time that don’t affect the operation of the software interface, Android background Service technology has been adopted. The communication of Activity and Service that is realized by BroadcastReceiver, has realized the run of users operation interface and background program in the same time, improve the interactivity and real time of the software. The experimental results show that the software has realized the local and remote intelligent measure and control for the plant growth cabinet.

Petiole Segmentation Method Based on Multi-Structure Elements Morphology

Leaf and petiole can be divided by using computer image processing technology. It has an important role for the feature extracting of leaf and the accuracy of characteristic analysis. In this paper, we compare the distribution of multi-channel histogram in HSV color model, and then select the appropriate channel and threshold value to process the image. Here we use the method of mathematical morphology with the element of multi-structure and multi-scales to smooth the leaf boundary and eliminate the isolated points in the image. Finally, we realize the effective segmentation between the leaf and petiole. Results show that when the condition of the color difference is not obvious, we can better remove the petiole by the use of saturation and brightness channel, and choose the appropriate structure elements according to the direction and size of the petiole. It can provide technical support for the accurate segmentation and extraction of feature information on the leaf.

Detection method of vegetable maturity based on neural network and bayesian information fusions

In order to better grasp the maturity of vegetables, this paper proposes a method which makes full use of external morphological characteristics of vegetables to infer the maturity of vegetables. Especially extracting the morphological features of the root and combine them with the ground morphological features. In this paper, firstly, vegetable images are disposed by threshold segmentation and feature extraction using the image processing toolbox of Matlab. Through this way, the value of leaf crown projected area, plant height, root length and root side area will be got. Secondly, Features of ground part and underground part can be used as training samples for corresponding neural network maturity detection models. Ultimately, Bayesian theory is utilized to process information fusion with obtained values of each neural network. The results show that this method improved the accuracy of detection.

Design of Hardware of Smart Plant Growth Cabinet

With the increase of nuclear radiation, soil pollution, water pollution and other man-made pollution, people are increasingly concerned about healthy eating. It gradually becomes a hot spot for scientists to develop a new green planting method. In this context, a design of smart plant growth cabinet based on LED lighting is put forward in this paper. From the analysis of social needs and the user′s analysis, transformation has been made from user requirement to the thinning mapping of product function and performance requirements. The main function and implementation principle of it and the design of the hardware are introduced in detail. This design has simple hardware structure, convenient maintenance, good stability, anti-interference ability and other characteristics. It has broad application prospect and great using value.

Simulated Annealing Algorithm Based Single Robot Odor Source Localization Strategy

This paper presents a search strategy for single mobile robots to realize the active olfaction (also called odor/gas source localization or plume tracing). The odor source localization is regarded as a kind of dynamic function optimization problem in this article, using the simulated annealing algorithm to calculate the optimal solution of density distribution function, namely the odor source location. The simulation experiments results in indoor ventilated environment show that the robot can track in plume and locate the odor source under the area of the 10m*10m, and it can effectively jump out of local maximum values in the process of search.

System Identification of Beating Heart Medical Surgery Assisted Robot

The wide bandwidth and high amplitude feature of the beating heart motion makes surgeon hard to achieve the coronary artery bypass graph surgery on his or her own. The robot could help to cancel the relative motion between the end effecter of the robotic tool and certain point on the heart surface. Therefore a stationary operation screen could be used for the surgeon. The precise robot modeling is the prerequisite for tracking control algorithm. In the paper, we process the input and output data by using the nonlinear optimization method to obtain the system model. Next, we discuss the system identification related problems and modify the model. Finally, a good system model is achieved.

Method for Non-Destructive Measurement of Leaf Area Based on Binocular Vision

Modern agriculture is developing towards the direction of intelligent. We can realize the nondestructive measurement of plant leaf area by using the digital camera. In this paper, we make the same ratio between the camera’s screen and the background plate to overcome the problem of geometric distortion. Then we use two perpendicular digital cameras from the front and side to collect images respectively for curved leaf. Because there are characteristics of the image grey value mutation on the rage of the vane, we can extract the leaf by image segmentation. The leaf area can be calculated by the statistic of the pixels number according to the projection principle. Experiments show that, the error of leaf area measurement reduces from 13.51% to 5.93% by binocular vision. So this method not only can get the measurement of leaf area data, but also can effectively avoid the two-dimensional image distortion and improve the accuracy of leaf area calculation.

The Mixed Kalman and H Infinity Filter Based Robust Model Following Control Algorithm for CABG Beating Heart Surgery

In the CABG surgery the robot dynamically cancels the relative motion between the point of interest (POI) on the beating heart and robotic instruments, such that the surgeon can operate as if the heart is stationary. However, the highly nonlinear and non-stationary nature of the beating heart motion poses difficulties for robot to follow the characteristics of the beating heart motion. Furthermore, the surgery has potential safety risk if the robot system could not track the POI properly. Therefore, in order to minimize the effects caused by the uncertainty in the heart motion model during the surgery, a robust prediction based model following control algorithm is proposed here. The adaptive Autoregressive (AR) model integrated the mixed Kalman and H infinity filter to estimate the state of the beating heart motion in the sense of minimizing minimize both RMS motion estimation error and worst case motion estimation error. In addition, the linear quadratic optimal tracking theory was used to implement the model following controller. In such way a complicated heart motion tracking problem transformed to dynamic model following problem and the robust property of the tracking control is more effective. The method is verified by two prerecorded distinguished datasets on 3D test bed robotics.

The Mixing Algorithm of ACO and GA based Global Path Planning Method for Mobile Robot

An ant colony algorithm is a stochastic searching optimization algorithm that is based on the heuristic behavior of the biologic colony. Its positive feedback and coordination make it possible to be applied to a distributed system. It has favorable adaptability in solving combinatorial optimization and has great development potential for its connotative parallel property. This study focused on global path planning with an ant colony algorithm in an environment based on grids, which explores a new path planning algorithm. How to present and update the pheromone of an ant system was investigated. The crossover operation of a genetic algorithm was used in the ant system for path optimization. Experimental results show that the algorithm has better path planning optimization ability than other algorithms.