Search

In this article, a novel model-free dynamic inversion-based Q-learning (DIQL) algorithm is proposed to solve the optimal tracking control (OTC) problem of unknown nonlinear input-affine discrete-time (DT) systems. Compared with the existing DIQL algorithm and the discount factor-based Q-learning (DFQL) algorithm, the proposed algorithm can eliminate the tracking error while ensuring that it is model-free and off-policy. First, a new deterministic Q-learning iterative scheme is presented, and based on this scheme, a model-based off-policy DIQL algorithm is designed. The advantage of this new scheme is that it can avoid the training of unusual data and improve data utilization, thereby saving computing resources. Simultaneously, the convergence and stability of the designed algorithm are analyzed, and the proof that adding probing noise into the behavior policy does not affect the convergence is presented. Then, by introducing neural networks (NNs), the model-free version of the designed algorithm is further proposed so that the OTC problem can be solved without any knowledge about the system dynamics. Finally, three simulation examples are given to demonstrate the effectiveness of the proposed algorithm.

Background: Osteoporosis is characterized by reduced bone mass and destruction of trabecular bone microstructure, increasing fracture risk. Newly marketed drugs have certain advantages in osteoporosis, but high health care expenditures and side effects limit their application. Therefore, it is urgent to explore new approaches to break the therapeutic deadlock in the current clinical condition. Myricitrin has many biological activities. However, the effect on antiosteoporosis has not fully been elucidated yet.Methods: The effect of myricitrin on the proliferation of immortalized bone marrow mesenchymal stem cells was evaluated using Cell Counting Kit-8 assays. Alizarin Red staining, Alkaline phosphatase staining, and activity were performed to elucidate osteogenesis. qRT-PCR and Western blot were performed to evaluate the expression of osteo-specific genes and proteins. To screen for candidate targets, sequencing of mRNA transcriptome genes was used through bioinformatics analysis. Western blot and Molecular docking analysis were used to examine target signaling markers. Rescue experiments were used to confirm myricitrin can enhance osteogenic differentiation of imBMSCs reversely. Moreover, ovariectomy mice were used to assess the effect of myricitrin on bone loss in vivo by Western blot, Micro-CT, Hematoxylin and Eosin staining, and immunochemistry.Results: Myricitrin can significantly enhance osteo-specific genes and proteins expression, alkaline phosphatase activity, and calcium deposition in imBMSCs in vitro. These phenomena were accompanied by an upregulation of the PI3K/AKT signaling pathway via bioinformatics analysis and Western blot. In addition, the improvement of osteogenesis due to myricitrin intervention was partially reversed by copanlisib, which is a PI3K inhibitor. Osteopontin (OPN), osteoprotegerin (OPG), p-PI3K, and p-AKT were significantly increased in the low dose and high dose myricitrin groups compared with the OVX group. Micro-CT, H&E staining, and immunochemistry showed that bone mass was significantly increased in low dose and high dose groups compared with the OVX group. Furthermore, myricitrin can attenuate bone loss in ovariectomy mice.Conclusion: These data suggest that myricitrin enhances the osteogenic differentiation of imBMSCs, partly via activation of the PI3K/AKT signaling pathway. Myricitrin may be a potential drug development for the treatment of postmenopausal osteoporosis.

Background: This study used original survey data to quantitatively investigate the associations between individuals’ perception of locally present wastewater pollution and their self-rated health. Methods: This research used the data from large-scale surveys covering all the 31 provinces and equivalent administrative units in mainland China and interviewed 6112 participants. The ordered logit method was employed to estimate the models. Results: The results indicated that individuals’ perceptions of local industrial and domestic wastewater pollution significantly decrease their self-rated health. If industrial wastewater pollution was reported, the possibility of the observers indicating lower levels of self-rated current health, comparing to the past year, and comparing with peers, all increased by 26% (p < 0.001), 23% (p = 0.005), and 18% (p = 0.006), respectively. Likewise, perceived domestic wastewater pollution led to the increase by 21% (p = 0.012), 17% (p = 0.034), and 33% (p = 0.000), respectively. Meanwhile, reported industrial wastewater pollution also has an obvious negative effect on individuals’ health performance, such as being more fatigued and upset. Conclusions: The survey clearly shows that Chinese individuals who are aware of water pollution in their living environment tend to experience more negative health outcomes, which adds additional urgency to improving wastewater treatment.

With more and more extensive research and application of wearable devices, their continuous working time has become a prominent problem which is increasingly concerned. Within the limitations of existing battery technology, environmental energy harvesting can be used to enhance the continuous working time of wearable devices. To solve this problem, this paper designs a collection system for heat energy and radiofrequency (RF) energy in the environment. This system designs a thermoelectric collector based on thermoelectric generation, an RF energy collector based on a rectifier antenna, and a mixed energy collection controller based on power trajectory tracking. This system can generate power according to the temperature difference between environment and human body and RF energy in the environment. This system uses the heart rate sampling algorithm to accurately collect the required voltage and current information while saving system energy consumption. The maximum power output algorithm based on power trajectory tracking keeps the maximum power output of the energy collection system at all times. In this paper, two kinds of environmental energy models and a mixed energy acquisition system model are established by MATLAB. After setting reasonable environmental conditions, we get the results to verify the rationality and effectiveness of this environment energy collection system.

Designing a robot with the best accuracy is always the attractive research direction in robot community. In order to create a Gough-Stewart platform with guaranteed accuracy performance for a dedicated controller, this paper describes a novel advanced optimal design methodology: control-based design methodology. This advanced optimal design method considers the controller positioning accuracy in the design process for getting the optimal geometric parameters of the robot. In this paper, three types of visual servoing controllers are applied to control the motions of the Gough-Stewart platform: leg-direction-based visual servoing, line-based visual servoing and image moment visual servoing. Depend on these controllers, the positioning error models considering the camera observation error together with the controller singularities are analyzed. In the next step, the optimization problems are formulated in order to get the optimal geometric parameters of the robot and the placement of the camera for the Gough-Stewart platform for each type of controller. Then, we perform the co-simulations on the three optimized Gough-Stewart platforms in order to test the positioning accuracy and the robustness with respect to the manufacturing errors. It turns out that the optimal control-based design methodology helps getting both the optimum design parameters of the robot and the performance of the controller {robot + dedicated controller}.

In this study, we propose a parallel processing method for analyzing video-image radiation-response signals and suppressing radiation noise. We studied the linear-representation law of various image-information components on the radiation dose rate. Subsequently, the simulation images were used to examine the response-signal extract and radiation-noise suppression. The results indicate that the majority of response signals in the global image comprise forward superposition. The peak signal-to-noise ratio of the red channel was significantly improved when the noise signal-substitution algorithm and median filter were applied successively. Real-time radiation dose-rate measurements and clear images under irradiation can be obtained simultaneously.

Background: Previous studies had compared the effectiveness of surgical and conservative treatment in proximal humeral fracture (PHF), suggesting that both treatments were effective for displaced fractures. However, which treatment is more effective, especially in the elderly patient population, remains controversial to date. This study aims to evaluate the efficacy of the two treatment methods, including functional outcome scores, complications, etc.Methods: We searched four databases (MEDLINE, Web of Science, EMBASE, Cochrane Library) from the inception of the databases to March 2020. Within the scope of the search, the publication time and language of the relevant literatures were not limited. The meta-analysis directly compared the results and complications of the two groups and subgroups. Results: We included 11 comparative studies into our meta-analysis. The contents of the study included 7 randomized controlled trials (RCTs) and 4 non-randomized controlled trials (nRCTs). The pooled data demonstrated that there was no significant difference in postoperative Constant score between surgical and conservative treatment in the RCT subgroup, but there were differences among the nRCT and prosthesis subgroup. Although there was no significant difference between the surgery and conversation group in the major postoperative complications, the subgroup analysis showed fewer complications in the joint prosthesis replacement group. Conclusion: Both treatments could provide better clinical results for patients. The joint prosthesis replacement may offer additional benefits in terms of reduced complications and postoperative function. Furthermore, due to the increasing number of prosthesis replacements and the advent of the aging age, we should consider the individual differences between patients and enrich the clinical efficacy through the personal experience of surgeons before choosing treatment options.

Background: Gait impairments including shuffling gait and hesitation are common in people with Parkinson's disease (PD), and have been linked to increased fall risk and freezing of gait. Nowadays the gait metrics mostly focus on the spatiotemporal characteristics of gait, but less is known of the angular characteristics of the gait, which may provide helpful information pertaining to the functional status and effects of the treatment in PD. Objective: This study aimed to quantify the angles of steps during walking, and explore if this novel step angle metric is associated with the severity of PD and the effects of the treatment including the acute levodopa challenge test (ALCT) and deep brain stimulation (DBS). Methods: A total of 18 participants with PD completed the walking test before and after the ALCT, and 25 participants with PD completed the test with the DBS on and off. The walking test was implemented under two conditions: walking normally at a preferred speed (single task) and walking while performing a cognitive serial subtraction task (dual task). A total of 17 age-matched participants without PD also completed this walking test. The angular velocity was measured using wearable sensors on each ankle, and three gait angular metrics were obtained, that is mean step angle, initial step angle, and last step angle. The conventional gait metrics (ie, step time and step number) were also calculated. Results: The results showed that compared to the control, the following three step angle metrics were significantly smaller in those with PD: mean step angle (F-1,F-48=69.75, P

Load is the main external disturbance of a parallel robot manipulator. This disturbance will cause dynamic coupling among different degrees of freedom and make heaps of model-based control methods difficult to apply. In order to compensate this disturbance, it is crucial to obtain an accurate dynamic model of load. However, in practice, the load is always uncertain and its dynamic parameters are arduous to know a priori. To cope with this problem, this paper proposes a novel and simple approach to identify the dynamic parameters of load. Firstly, the dynamic model of the parallel robot manipulator with uncertain load is established and the dynamic coupling caused by load is also analyzed. Then, according to the dynamic model, the excitation signal is designed and a weak nonlinear dynamic model is derived. Furthermore, the identification model is presented and the identification algorithm based on the extended Kalman filter is designed. Lastly, numerical simulation results, obtained using a six-degree-of-freedom Gough–Stewart parallel manipulator, demonstrate the good estimation performance of the proposed method.

In order to realize hardware-in-the-loop simulation of commonly used power electronic circuits, a method of joint simulation using Simcape and field programmable gate array (FPGA) is proposed. This article adopts a switch modeling method suitable for high-frequency characteristics, namely the switch function method. It effectively solves the problem of rigidity non-convergence in the modeling of the binary resistance method, as well as the transient error caused by the binary LC method when the system state is switched with the increase of the switching frequency, which reduces the simulation accuracy. Taking the single-phase full-bridge inverter circuit as an example, a Simscape model based on the state-space averaging method was established, and the HDL code was automatically generated using the Simscape HDL Workflow Advisor tool, and the hardware-in-the-loop simulation was completed with FPGA. Finally, the simulation results with Simulink comparison verify the effectiveness and feasibility of the method.

We present γ-ray radiation detection in a neutron radiation environment using a monolithic active pixel sensor (MAPS) camera without conversion or shielding layers. The measured output signal is the sum of the pedestal value, noise, and real radiation response signal. The sensor response shows that the MAPS camera is sensitive to neutrons and can capture a single photon. The number of pixels with a signal exceeding 100 exhibits a strong dependence on the dose rate and is the best indicator of this value. Therefore, a MAPS camera can be efficiently used as a radiation detection sensor in a robotic system, further limiting human errors in performing radiation detection in complex nuclear radiation environments.

BACKGROUND Gait impairments including shuffling gait and hesitation are common in people with Parkinson’s disease (PD), and have been linked to increased fall risk and freezing of gait. Nowadays the gait metrics mostly focus on the spatiotemporal characteristics of gait, but less is known of the angular characteristics of the gait, which may provide helpful information pertaining to the functional status and effects of the treatment in PD. OBJECTIVE This study aimed to quantify the angles of steps during walking, and explore if this novel step angle metric is associated with the severity of PD and the effects of the treatment including the acute levodopa challenge test (ALCT) and deep brain stimulation (DBS). METHODS A total of 18 participants with PD completed the walking test before and after the ALCT, and 25 participants with PD completed the test with the DBS on and off. The walking test was implemented under two conditions: walking normally at a preferred speed (single task) and walking while performing a cognitive serial subtraction task (dual task). A total of 17 age-matched participants without PD also completed this walking test. The angular velocity was measured using wearable sensors on each ankle, and three gait angular metrics were obtained, that is mean step angle, initial step angle, and last step angle. The conventional gait metrics (ie, step time and step number) were also calculated. RESULTS The results showed that compared to the control, the following three step angle metrics were significantly smaller in those with PD: mean step angle (F1,48=69.75, PF1,48=15.56, PF1,48=61.99, PF1,38=5.77, P=.02, partial eta-square=0.13; DBS: F1,52=8.53, P=.005, partial eta-square=0.14) and last step angles (ACLT: F1,38=10, P=.003, partial eta-square=0.21; DBS: F1,52=4.96, P=.003, partial eta-square=0.09), but no significant changes were observed in step time and number after the treatments. Additionally, these step angles were correlated with the Unified Parkinson's Disease Rating Scale, Part III score: mean step angle (single task: r=–0.60, Pr=–0.52, Pr=–0.35, P=.006; dual task: r=–0.35, P=.01), and last step angle (single task: r=–0.43, P=.001; dual task: r=–0.41, P=.002). CONCLUSIONS This pilot study demonstrated that the gait angular characteristics, as quantified by the step angles, were sensitive to the disease severity of PD and, more importantly, can capture the effects of treatments on the gait, while the traditional metrics cannot. This indicates that these metrics may serve as novel markers to help the assessment of gait in those with PD as well as the rehabilitation of this vulnerable cohort.

DeepLab v3+ neural network shows excellent performance in semantic segmentation. In this paper, we proposed a segmentation framework based on DeepLab v3+ neural network and applied it to the problem of hyperspectral imagery classification (HSIC). The dimensionality reduction of the hyperspectral image is performed using principal component analysis (PCA). DeepLab v3+ is used to extract spatial features, and those are fused with spectral features. A support vector machine (SVM) classifier is used for fitting and classification. Experimental results show that the framework proposed in this paper outperforms most traditional machine learning algorithms and deep-learning algorithms in hyperspectral imagery classification tasks.

Membrane gas absorption (MGA) based on ceramic membranes is considered to be an effective technology for removing SOx from ship exhaust. To develop a convenient method for designing and amplifying a membrane contactor, the critical gas velocity was defined. Two simple connection modes, series and parallel modules, were assembled and compared with a single-tube membrane contactor. The critical gas velocity is associated with the size of the membrane element and module, based on which an empirical equation can be formulated to optimize the gas velocity in different membrane contactors. In addition, another amplifying method, using multiple tubes in one shell, was also studied with computational fluid dynamics (CFD) simulations. The results imply that the addition of baffles improves the desulfurization performance of the membrane contactor at the critical gas velocity. The methods and conclusions presented herein provide a design basis for using the contactor at an industrial scale.

In wireless local area networks (WLANs), link-layer multicast is a promising technology for many multimedia applications, e.g., video streaming, as multicast frames can reach multiple clients simultaneously. However, the efficiency of multicast in WLANs is usually low since multicast frames are transmitted at a basic data rate to reach clients with poor channel quality. Moreover, the reliability of multicast cannot be guaranteed either, as multicast transmissions are not acknowledged. Some recent works have utilized smart antennas to improve multicast performance. However, most of them require customized hardware and are not designed for the latest IEEE 802.11 standard, 802.11n WLANs. More importantly, these works did not consider the impact of AP association strategies on the performance of link-layer multicast. In this paper, we study the problem of link-layer multicast in large-scale 802.11n WLANs where each AP is equipped with a smart antenna that supports multiple antenna patterns. Based on channel gains of various antenna patterns from different APs, we choose the associated AP for multicast clients, partition associated clients of each AP into multiple groups, select an antenna pattern and data rate for each group of client, and transmit multicast frames to each group once. We first examine the limitation of multicast in 802.11n WLANs and the benefits of smart antennas to multicast via experiments. We then introduce the system model and formulate the problem into an optimization problem, and prove its NP-hardness. The objective is to minimize the time to transmit a multicast frame to all clients while guaranteeing high packet reception ratio (PRR) for each client. For generic WLANs where APs are sparsely deployed, we propose an optimal algorithm under the condition that the PRR of antenna patterns and data rates are known for every client. For generic large-scale WLANs, we further propose an on-line algorithm in which the AP association strategy, the partition of clients, the antenna pattern and data rate of each group are adapted dynamically, based on PRR reports from clients. We have implemented the on-line algorithm on off-the-shelf WLAN products, and conducted extensive experiments and simulations to evaluate the performance. The results show that the proposed algorithm can significantly improve multicast performance compared to other schemes, and at the same time guarantee high PRR for all clients.

A considerable number of studies have been conducted to study the microbial profiles in inflammatory conditions. A common phenomenon in inflammatory bowel disease (IBD) is the reduction of the diversity of microbiota, which demonstrates that microbial diversity negatively correlates with disease severity in IBD. Increased microbial diversity is known to occur in disease remission. Species diversity plays an important role in maintaining the stability of the intestinal ecosystem as well as normal ecological function. A reduction in microbial diversity corresponds to a decrease in the stability of the ecosystem and can impair ecological function. Fecal microbiota transplantation (FMT), probiotics, and prebiotics, which aim to modulate the microbiota and restore its normal diversity, have been shown to be clinically efficacious. In this study, we hypothesized that a reduction in microbial diversity could play a role in the development of IBD.

This paper proposes an improved ant colony algorithm to achieve efficient searching capabilities of path planning in complicated maps for mobile robot. The improved ant colony algorithm uses the characteristics of A

Recent studies have shown that significant benefit can be achieved in wireless sensor networks (WSNs) by employing mobile collectors for data gathering via short-range communications. A typical scenario for such a scheme is that a mobile collector roams over the sensing field and pauses at some anchor points on its moving tour such that it can traverse the transmission range of all the sensors in the field and directly collect data from each sensor. In this paper, we study the performance optimization of such mobile data gathering by formulating it into a cost minimization problem constrained by channel capacity, required minimum data uploads from each sensor and bound of total sojourn time at all anchor points. In order to provide an efficient and distributed algorithm, we decompose this global optimization problem into two subproblems that can be solved by each sensor and the mobile collector, respectively. We show that such decomposition can be characterized as a pricing mechanism, in which each sensor independently adjusts its payment for the data uploading opportunity based on the shadow prices of different anchor points. Correspondingly, we give an efficient algorithm to jointly solve the two subproblems. Our theoretical analysis demonstrates that the proposed algorithm can achieve the optimal data control for each sensor and the optimal sojourn time allocation for the mobile collector, which minimizes the overall network cost. Finally, extensive simulation results further validate that our algorithm achieves lower cost than the compared data gathering strategy.

One of the motivations of OptoBot Lab is to train primer students into qualified engineers or researchers. The series training programs have been designed by supervisors and implemented with tutoring for students to test and practice their knowledge from textbooks. An environment perception experiment using a 32 layers laser scanner is described in this paper. The training program design and laboratory operation is introduced. The four parts of the experiments which are preparation, sensor calibration, 3D space reconstruction, and object recognition, are the participating students’ main tasks for different teams. This entire program is one of the series training programs that play significant role in establishing solid research skill foundation for opto-electronic students.

Exclusive enteral nutrition (EEN) has been proven to be effective and safe in treating pediatric Crohn’s disease (CD). EEN induces pediatric CD remission possibly through three pathways: (1) direct anti-inflammatory effects, (2) improved epithelial barrier function, and (3) modulation of the gut microbiota. Recent studies have demonstrated that modulation of the gut microbiota plays a major role in EEN-induced remission. Variations of microbial components, which directly influence the diversity and metabolic functions of the gut microbiota, are closely associated with the immunological conditions of the gut and the susceptibility to diseases. The reduction of proinflammatory microbial components and harmful microbial metabolites after EEN treatment greatly decreases the inflammatory injuries of the gut.

This study proposes an adaptive neural network controller for a 3-DOF robotic manipulator that is subject to backlash-like hysteresis and friction. Two neural networks are used to approximate the dynamics and the hysteresis non-linearity. A neural network, which utilises a radial basis function approximates the robot's dynamics. The other neural network, which employs a hyperbolic tangent activation function, is used to approximate the unknown backlash-like hysteresis. The authors also consider two cases: full state and output feedback control. For output feedback, where system states are unknown, a high gain observer is employed to estimate the states. The proposed controllers ensure the boundedness of the control signals. Simulations are also performed to show the effectiveness of the controllers.

Recently, several multi-channel MAC protocols have been proposed for wireless sensor networks (WSNs) to improve network capacity and boost energy efficiency. In addition, cooperative multiple-input multiple-output (MIMO) technique has been shown to be able to significantly enhance the energy efficiency of WSNs if properly configured. However, these two promising techniques have not been jointly considered in clustered WSNs. In this paper, we explore such a joint design by proposing a novel MAC protocol for clustered WSNs that takes advantage of both multiple channels and cooperative MIMO to further boost network performance. Specifically, sensor nodes in a WSN are organized into clusters and each cluster head selects some cooperative nodes to help forward traffic to or receive from other clusters by utilizing a cooperative MIMO technique. For intra-cluster communications, different channels are assigned to adjacent clusters to reduce collisions, while for inter-cluster communications, cooperative MIMO links are scheduled to improve energy efficiency and concurrent transmissions are enabled by assigning different channels to them. We also provide a network-wide time synchronization approach to facilitating the functionality of this protocol. We carry out extensive simulations and the results demonstrate that the proposed protocol can significantly increase throughput and improve energy efficiency compared to other schemes. For example, for a WSN with 150 sensors in a 250 i? 250 m 2 field, with five available channels, the protocol can achieve three times saturated throughput while saving 14 % energy per bit for inter-cluster communications. Propose a multi-channel, cooperative MIMO MAC protocol for clustered WSNs.Employ cooperative MIMO links for inter-cluster communications.Assign orthogonal channels to interfering cooperative links to boost throughput.Numerical results show that energy efficiency of WSNs is remarkably improved.Simulation results show that network throughput is significantly enhanced.

As the latest amendment of IEEE 802.11 standard, 802.11n allows a maximum raw data rate as high as 600 Mbps, making it a desirable candidate for wireless local area network (WLAN) deployment. In typical deployment, the coverage areas of nearby access points (APs) usually overlap with each other to provide satisfactory coverage and seamless mobility support. Clients tend to associate (connect) to the AP with the strongest signal strength, which may lead to poor client throughput and overloaded APs. Although a number of AP association schemes have been proposed for IEEE 802.11 WLANs in the literature, the challenges brought by the new features in 802.11n have not been thoroughly studied nor the impact of legacy 802.11a/b/g clients in 802.11n WLANS on AP association. To fill in this gap, in this paper, we explore AP association for 802.11n with heterogeneous clients (802.11a/b/g/n). We first present a bi-dimensional Markov model to estimate the uplink and downlink throughput of clients and formulate AP association into an optimization problem, aiming at providing each client a bandwidth proportional to its usable data rate. Based on this Markov model, we propose an on-line AP association algorithm under the condition that each client can acquire timely information of all clients associated with nearby APs. Furthermore, for WLANs with densely deployed APs, we provide another on-line AP association algorithm with lower complexity, which takes full advantage of 802.11n transmissions by simply associating different types of clients with different APs. We have conducted extensive simulations and experiments to validate the proposed algorithms. The results show that our algorithms can significantly improve both 802.11n throughput and aggregated network throughput under various network scenarios, compared to previous AP association schemes. Our experiments also confirm the effectiveness of the algorithms in enhancing network throughput, maintaining proportional fairness among clients, and balancing load among APs.

The MEMS gyroscope zero drifts dominantly affects measurement accuracy of gyroscope. For the nonlinear problem that MEMS gyro zero drift varying with the temperature changes, the wavelet transform was performed to filter the static experiment zero drift data, then the zero drift was investigate by combing with the improved grey prediction theory. Finally a temperature compensation model based on wavelet transform and the improved grey prediction was obtained. Comparing with conventional compensation model algorithm, our model has smaller root mean square error and mean absolute error, the drift of root mean square error and mean absolute error of the MEMS gyroscope were reduced to 0.0250 and 0.0180 respectively. The temperature compensation system has good potential of theoretical and engineering application to improve the gyro accuracy of measurement.

As the latest IEEE 802.11 standard, 802.11n applies several new technologies, such as multiple input multiple output (MIMO), channel bonding, and frame aggregation to greatly improve the rate, range and reliability of wireless local area networks (WLANs). In 802.11n WLANs, access points (APs) are often densely deployed to provide satisfactory coverage. Thus nearby APs should operate at non-overlapping channels to avoid mutual interference. It is challenging to assign channels in legacy 802.11a/b/g WLANs due to the limited number of channels. Channel assignment becomes more complex in 802.11n WLANs, as the channel bonding in 802.11n allows WLAN stations (APs and clients) to combine two adjacent, non-overlapping 20MHz channels together for transmission. On the other hand, IEEE 802.11n is backward compatible, such that 802.11n clients will coexist with legacy clients in 802.11n WLANs. Legacy clients may affect the performance of nearby 802.11n clients, and reduce the effectiveness of channel bonding. Based on these observations, in this paper, we study channel assignment in 802.11n WLANs with heterogeneous clients. We first present the network model, interference model, and throughput estimation model to estimate the throughput of each client. We then formulate the channel assignment problem into an optimization problem, with the objective of maximizing overall network throughput. Since the problem is NP-hard, we give a distributed channel assignment algorithm based on the throughput estimation model. We then present another channel assignment algorithm with lower complexity, and aim at minimizing interference experienced by high-rate, 802.11n clients. We have carried out extensive simulations to evaluate the proposed algorithms. Simulation results show that our algorithms can significantly improve the network throughput of 802.11n WLANs, compared with other channel assignment algorithms.