Hardware implementation of memristor-based artificial neural networks.

Artificial Intelligence (AI) is currently experiencing a bloom driven by deep learning (DL) techniques, which rely on networks of connected simple computing units operating in parallel. The low communication bandwidth between memory and processing units in conventional von Neumann machines does not support the requirements of emerging applications that rely extensively on large sets of data. More recent computing paradigms, such as high parallelization and near-memory computing, help alleviate the data communication bottleneck to some extent, but paradigm- shifting concepts are required. Memristors, a novel beyond-complementary metal-oxide-semiconductor (CMOS) technology, are a promising choice for memory devices due to their unique intrinsic device-level properties, enabling both storing and computing with a small, massively-parallel footprint at low power. Theoretically, this directly translates to a major boost in energy efficiency and computational throughput, but various practical challenges remain. In this work we review the latest efforts for achieving hardware-based memristive artificial neural networks (ANNs), describing with detail the working principia of each block and the different design alternatives with their own advantages and disadvantages, as well as the tools required for accurate estimation of performance metrics. Ultimately, we aim to provide a comprehensive protocol of the materials and methods involved in memristive neural networks to those aiming to start working in this field and the experts looking for a holistic approach.

CircEYA3 aggravates intervertebral disc degeneration through the miR-196a-5p/EBF1 axis and NF-κB signaling.

Intervertebral disc degeneration (IDD) is a well-established cause of disability, and extensive evidence has identified the important role played by regulatory noncoding RNAs, specifically circular RNAs (circRNAs) and microRNAs (miRNAs), in the progression of IDD. To elucidate the molecular mechanism underlying IDD, we established a circRNA/miRNA/mRNA network in IDD through standardized analyses of all expression matrices. Our studies confirmed the differential expression of the transcription factors early B-cell factor 1 (EBF1), circEYA3, and miR-196a-5p in the nucleus pulposus (NP) tissues of controls and IDD patients. Cell proliferation, apoptosis, and extracellular mechanisms of degradation in NP cells (NPC) are mediated by circEYA3. MiR-196a-5p is a direct target of circEYA3 and EBF1. Functional analysis showed that miR-196a-5p reversed the effects of circEYA3 and EBF1 on ECM degradation, apoptosis, and proliferation in NPCs. EBF1 regulates the nuclear factor kappa beta (NF-кB) signalling pathway by activating the IKKß promoter region. This study demonstrates that circEYA3 plays an important role in exacerbating the progression of IDD by modulating the NF-κB signalling pathway through regulation of the miR196a-5p/EBF1 axis. Consequently, a novel molecular mechanism underlying IDD development was elucidated, thereby identifying a potential therapeutic target for future exploration.

Stochastic neuro-fuzzy system implemented in memristor crossbar arrays.

Neuro-symbolic artificial intelligence has garnered considerable attention amid increasing industry demands for high-performance neural networks that are interpretable and adaptable to previously unknown problem domains with minimal reconfiguration. However, implementing neuro-symbolic hardware is challenging due to the complexity in symbolic knowledge representation and calculation. We experimentally demonstrated a memristor-based neuro-fuzzy hardware based on TiN/TaOx/HfOx/TiN chips that is superior to its silicon-based counterpart in terms of throughput and energy efficiency by using array topological structure for knowledge representation and physical laws for computing. Intrinsic memristor variability is fully exploited to increase robustness in knowledge representation. A hybrid in situ training strategy is proposed for error minimizing in training. The hardware adapts easier to a previously unknown environment, achieving ~6.6 times faster convergence and ~6 times lower error than deep learning. The hardware energy efficiency is over two orders of magnitude greater than field-programmable gate arrays. This research greatly extends the capability of memristor-based neuromorphic computing systems in artificial intelligence.

An Ultrasensitive Biomimetic Optic Afferent Nervous System with Circadian Learnability.

The optic afferent nervous system (OANS) plays a significant role in generating vision and circadian behaviors based on light detection and signals from the endocrine system. However, the bionic simulation of this photochemically mediated behavior is still a challenge for neuromorphic devices. Herein, stimuli of neurotransmitters at ultralow concentrations and illumination are coupled to artificial synapses with the aid of biofunctionalized heterojunction and tunneling to successfully simulate a circadian neural response. Furthermore, the mechanisms underlying the photosensitive synaptic current in response to stimuli are described. Interestingly, this OANS is demonstrated to be capable of mimicking normal and abnormal circadian learnability by combining the measured synaptic current with a three-layer spike neural network. Strong theoretical and experimental evidence, as well as applications, are provided for the proposed biomimetic OANS to demonstrate that it can reproduce biological circadian behavior, thus establishing it as a promising candidate for future neuromorphic intelligent robots.

Fibrous MXene Synapse-Based Biomimetic Tactile Nervous System for Multimodal Perception and Memory.

Biomimetic tactile nervous system (BTNS) inspired by organisms has motivated extensive attention in wearable fields due to its biological similarity, low power consumption, and perception-memory integration. Though many works about planar-shape BTNS are developed, few researches could be found in the field of fibrous BTNS (FBTNS) which is superior in terms of strong flexibility, weavability, and high-density integration. Herein, a FBTNS with multimodal sensibility and memory is proposed, by fusing the fibrous poly lactic acid (PLA)/Ag/MXene/Pt artificial synapse and MXene/EMIMBF4 ionic conductive elastomer. The proposed FBTNS can successfully perceive external stimuli and generate synaptic responses. It also exhibits a short response time (23 ms) and low set power consumption (17 nW). Additionally, the proposed device demonstrates outstanding synaptic plasticity under both mechanical and electrical stimuli, which can simulate the memory function. Simultaneously, the fibrous devices are embedded into textiles to construct tactile arrays, by which biomimetic tactile perception and temporary memory functions are successfully implemented. This work demonstrates the as-prepared FBTNS can generate biomimetic synaptic signals to serve as artificial feeling signals, it is thought that it could offer a fabric electronic unit integrating with perception and memory for Human-Computer interaction, and has great potential to build lightweight and comfortable Brain-Computer interfaces.

High-Performance Neuromorphic Computing and Logic Operation Based on a Self-Assembled Vertically Aligned Nanocomposite SrTiO3:MgO Film Memristor.

Neuromorphic computing based on memristors capable of in-memory computing is promising to break the energy and efficiency bottleneck of well-known von Neumann architectures. However, unstable and nonlinear conductance updates compromise the recognition accuracy and block the integration of neural network hardware. To this end, we present a highly stable memristor with self-assembled vertically aligned nanocomposite (VAN) SrTiO3:MgO films that achieve excellent resistive switching with low set/reset voltage variability (4.7%/-5.6%) and highly linear conductivity variation (nonlinearity = 0.34) by spatially limiting the conductive channels at the vertical interfaces. Various synaptic behaviors are simulated by continuously modulating the conductance. Especially, convolutional image processing using diverse crossbar kernels is demonstrated, and the artificial neural network achieves an overwhelming recognition accuracy of up to 97.50% for handwritten digits. Even under the perturbation of Poisson noise (λ = 10), 6% Salt and Pepper noise, and 5% Gaussian noise, the high recognition accuracies are retained at 95.43%, 94.56%, and 95.97%, respectively. Importantly, the logic memory function is proven experimentally based on the nonvolatile properties. This work provides a material system and design idea to achieve high-performance neuromorphic computing and logic operation.

A bidirectional thermal sensory leaky integrate-and-fire (LIF) neuron model based on bipolar NbOx volatile threshold devices with ultra-low operating current.

Brain-like artificial intelligence (AI) will become the main form and important platform in future computing. It will play an important and unique role in simulating brain functions, efficiently implementing AI algorithms, and improving computing power. Developing artificial neurons that can send facilitation/depression signals to artificial synapses, sense, and process temperature information is of great significance for achieving more efficient and compact brain-like computing systems. Herein, we have constructed a NbOx bipolar volatile threshold memristor, which could be operated by 1 µA ultra-low current and up to ∼104 switching ratios. By using a leaky integrate-and-fire (LIF) artificial neuron model, a bipolar LIF artificial neuron is constructed, which can realize the conventional threshold-driven firing, all-or-nothing spiking, refractory periods, and intensity-modulated frequency response bidirectionally at the positive/negative voltage stimulation, which will give the artificial synapse facilitation/depression signals. Furthermore, this bipolar LIF neuron can also explore different temperatures to output different signals, which could be constructed as a more compact thermal sensory neuron to avoid external harm to artificial robots. This study is of great significance for improving the computational efficiency of the system more effectively, achieving high integration density and low energy consumption artificial neural networks to meet the needs of brain-like neural computing.

Superlow Power Consumption Memristor Based on Borphyrin-Deoxyribonucleic Acid Composite Films as Artificial Synapse for Neuromorphic Computing.

Memristor synapses based on green and pollution-free organic materials are expected to facilitate biorealistic neuromorphic computing and to be an important step toward the next generation of green electronics. Metalloporphyrin is an organic compound that widely exists in nature with good biocompatibility and stable chemical properties, and has already been used to fabricate memristors. However, the application of metalloporphyrin-based memristors as synaptic devices still faces challenges, such as realizing a high switching ratio, low power consumption, and bidirectional conductance modulation. We developed a memristor that improves the resistive switching (RS) characteristics of Zn(II)meso-tetra(4-carboxyphenyl) porphine (ZnTCPP) by combining it with deoxyribonucleic acid (DNA) in a composite film. The as-fabricated ZnTCPP-DNA-based device showed excellent RS memory characteristics with a sufficiently high switching ratio of up to ∼104, super low power consumption of ∼39.56 nW, good cycling stability, and data retention capability. Moreover, bidirectional conductance modulation of the ZnTCPP-DNA-based device can be controlled by modulating the amplitudes, durations, and intervals of positive and negative pulses. The ZnTCPP-DNA-based device was used to successfully simulate a series of synaptic functions including long-term potentiation, long-term depression, spike time-dependent plasticity, paired-pulse facilitation, excitatory postsynaptic current, and human learning behavior, which demonstrates its potential applicability to neuromorphic devices. A two-layer artificial neural network was used to demonstrate the digit recognition ability of the ZnTCPP-DNA-based device, which reached 97.22% after 100 training iterations. These results create a new avenue for the research and development of green electronics and have major implications for green low-power neuromorphic computing in the future.

Strain engineering on electronic structure, effective mass and charge carrier mobility in monolayer YBr3.

In recent years, two-dimensional materials have significant prospects for applications in nanoelectronic devices due to their unique physical properties. In this paper, the strain effect on the electronic structure, effective mass, and charge carrier mobility of monolayer yttrium bromide (YBr3) is systematically investigated using first-principles calculation based on density functional theory. It is found that the monolayer YBr3undergoes energy band gap reduction under the increasing compressive strain. The effective mass and charge carrier mobility can be effectively tuned by the applied compressive strain. Under the uniaxial compressive strain along the zigzag direction, the hole effective mass in the zigzag direction (mao1_h) can decrease from 1.64m0to 0.45m0. In addition, when the uniaxial compressive strain is applied, the electron and hole mobility can up to ∼103cm2V-1s-1. The present investigations emphasize that monolayer YBr3is expected to be a candidate material for the preparation of new high-performance nanoelectronic devices by strain engineering.

Memristors based on NdNiO3 nanocrystals film as sensory neurons for neuromorphic computing.

By mimicking the behavior of the human brain, artificial neural systems offer the possibility to further improve computing efficiency and solve the von Neumann bottleneck. In particular, neural systems with perceptual capability expand the application field and lay a good foundation for the construction of perceptual storage and computational systems. However, research on neurons with perceptual functions is still relatively scarce, with most works focusing on optoelectronic synapses. The neuron is important for neuromorphic computing systems because neurons output excitatory or inhibitory stimuli to regulate the weight of synapses. Therefore, the construction of sensory neurons is crucial to expand the application range of brain-like neural computing. Here, an artificial sensory neuron is proposed, which is constructed using a photosensitive bipolar threshold switching memristor based on NdNiO3 (NNO) nanocrystals. These metallic phase nanocrystals can not only enhance the local electric field, but also act as a reservoir for defects (VoS) to guide the growth of conductive filaments and stabilize the performance of the device. They present stable bipolar threshold switching behavior with a low 120 nW set power, and the operating voltages decreased in light due to photocarrier action. A leaky integrate firing (LIF) neuron has been realized, which achieved key biological neuron functions, such as all-or-nothing spiking, threshold-driven firing, refractory period, and spiking frequency modulation. The LIF neurons receiving optical inputs have the properties of an artificial sensory neuron. It could regulate the spiking output frequency at different light densities, which could be used for a ship approaching a port. This work provides a promising hardware implementation towards constructing high-performance artificial intelligence to assist ships at night in a sensory system.

Silicon based Bi0.9La0.1FeO3 ferroelectric tunnel junction memristor for convolutional neural network application.

Computing in memory (CIM) based on memristors is expected to completely solve the dilemma caused by von Neumann architecture. However, the performance of memristors based on traditional conductive filament mechanism is unstable. In this study, we report a nonvolatile high-performance memristor based on ferroelectric tunnel junction (FTJ) Pd/Bi0.9La0.1FeO3 (6.9 nm) (BLFO)/La0.67Sr0.33MnO3 (LSMO) on a silicon substrate. The conductance of this device was adjusted by different pulse stimulation parameter to achieve various synaptic functions because of ferroelectric polarization reversal. Based on the multiple conductance characteristics of the devices and the high linearity and symmetry of weight updating, image processing and VGG8 convolutional neural network (CNN) simulation based on the devices were realized. Excellent results of the image processing are demonstrated. The recognition accuracy of CNN offline learning reached an astonishing 92.07% based on Cifar-10 dataset. This provides a more feasible solution to break through the bottleneck of von Neumann architecture.

Organic Passivation of Deep Defects in Cu(In,Ga)Se2 Film for Geometry-Simplified Compound Solar Cells.

Diverse defects in copper indium gallium diselenide solar cells cause nonradiative recombination losses and impair device performance. Here, an organic passivation scheme for surface and grain boundary defects is reported, which employs an organic passivation agent to infiltrate the copper indium gallium diselenide thin films. A transparent conductive passivating (TCP) film is then developed by incorporating metal nanowires into the organic polymer and used in solar cells. The TCP films have a transmittance of more than 90% in the visible and nearinfrared spectra and a sheet resistance of ~10.5 Ω/sq. This leads to improvements in the open-circuit voltage and the efficiency of the organic passivated solar cells compared with control cells and paves the way for novel approaches to copper indium gallium diselenide defect passivation and possibly other compound solar cells.

A SmNiO3 memristor with artificial synapse function properties and the implementation of Boolean logic circuits.

Recently, with the improvement of the requirements for fast and efficient data processing in the era of artificial intelligence, new forms of computing have come into being. Developing memristor devices that can simulate the brain's computing neutral network is particularly important for applications in the field of artificial intelligence. However, there are still some challenges in their biological function simulation and related circuit design. In this work, a memristor based on perovskite rare earth nickelates (RNiO3) is presented with excellent electrical performance, including three orders of magnitude higher current switching ratio and good repeatability, and can achieve bidirectional conductance regulation like weight modulation in bio-synapse. Furthermore, the synaptic like characteristics of the device have been mimicked successfully, such as excitatory postsynaptic current (EPSC), paired pulse facilitation (PPF), classical double pulse spike time-dependent plasticity (classical pair-STDP), triplet spike time-dependent plasticity (triplet-STDP), short-term plasticity (STP), long-term plasticity (LTP), the refractory period phenomenon and learning and forgetting rules. In particular, two synaptic devices and a leaky integrate-and-fire (LIF) neuron device are used to achieve a logic gate circuit to realize "AND", "OR", and "NOT" functions. The device paves the way for the application of high-density circuits in artificial intelligence.

Interface-engineered ferroelectricity of epitaxial Hf0.5Zr0.5O2 thin films.

Ferroelectric hafnia-based thin films have attracted intense attention due to their compatibility with complementary metal-oxide-semiconductor technology. However, the ferroelectric orthorhombic phase is thermodynamically metastable. Various efforts have been made to stabilize the ferroelectric orthorhombic phase of hafnia-based films such as controlling the growth kinetics and mechanical confinement. Here, we demonstrate a key interface engineering strategy to stabilize and enhance the ferroelectric orthorhombic phase of the Hf0.5Zr0.5O2 thin film by deliberately controlling the termination of the bottom La0.67Sr0.33MnO3 layer. We find that the Hf0.5Zr0.5O2 films on the MnO2-terminated La0.67Sr0.33MnO3 have more ferroelectric orthorhombic phase than those on the LaSrO-terminated La0.67Sr0.33MnO3, while with no wake-up effect. Even though the Hf0.5Zr0.5O2 thickness is as thin as 1.5 nm, the clear ferroelectric orthorhombic (111) orientation is observed on the MnO2 termination. Our transmission electron microscopy characterization and theoretical modelling reveal that reconstruction at the Hf0.5Zr0.5O2/ La0.67Sr0.33MnO3 interface and hole doping of the Hf0.5Zr0.5O2 layer resulting from the MnO2 interface termination are responsible for the stabilization of the metastable ferroelectric phase of Hf0.5Zr0.5O2. We anticipate that these results will inspire further studies of interface-engineered hafnia-based systems.

Effect of Anterior Cervical Decompression Fusion and Partial Resection of Uncinate Vertebra Joint on Cervical Sagittal Sequence in Patients with Non-Single-Segment Radiculopathy and its Correlation with Curative Effect: A Retrospective Analysis.

OBJECTIVE: Analyze the effect of preservation or resection of the partial uncinate joint on the sagittal sequence of the cervical vertebrae in patients with non-single-segment radiculopathy and the correlation between the sagittal sequence of the cervical vertebrae and the long-term effect after surgery, we explored whether it is necessary to perform partial resection of the uncinate joint in patients with cervical spondylotic radiculopathy undergoing anterior cervical decompression and fusion (ACDF). METHODS: The study retrospectively analyzed 96 patients with cervical spondylotic radiculopathy with more than two segments from August 2016 to January 2021, who underwent ACDF (ACDF group, 45 patients) or ACDF combined with partial uncinate joint resection (ACDF + UT group, 51 patients). Partial resection of the uncinate joint indicated removal of part of the uncinate joint and osteophyte based on the compression of the nerve root during surgery, whereas the uncinate joints in the ACDF group were retained completely. The imaging data and functional scores of the two groups were recorded before surgery, 1 month after surgery, and at the last follow-up. A paired t-test or rank sum test was applied to analyze the data. In addition, the correlation between the imaging parameters and functional scores was validated using the Pearson's test. RESULTS: All 96 patients successfully completed the surgery and were followed up for at least 12 months, with an average follow-up time of 14 months. At the last follow-up, the pain visual analog scale (VAS), neck disability index (NDI), and neck pain and disability scale (NPAD) scores of the two groups were significantly lower than those before surgery, and the Japanese Orthopaedic Association (JOA) score was significantly higher than that before surgery. At the last follow-up, compared with the ACDF+UT group, the NDI and NPAD scores in the ACDF group decreased more significantly (p < 0.05), and C2-7SVA, △C2-7SVA (the difference between C2-7 SVA at last follow-up and before operation), and T1S values decreased significantly (p < 0.05). The C2-7 Cobb angle was positively correlated with the JOA score and T1S (p < 0.05) and negatively correlated with the VAS, NDI, and NPAD scores and CGH-C7SVA (p < 0.05). C2-7SVA was positively correlated with CGH-C7SVA and T1S (p < 0.05). CONCLUSION: Patients with non-single-segmental cervical spondylotic radiculopathy and ACDF with or without uncinate joint resection can have effective improvement in the clinical effect and sagittal balance; however, partial uncinate joint resection has a certain negative impact on the long-term reconstruction of sagittal balance and long-term effects in patients after surgery.

Ligand-Triggered Self-Assembly of Flexible Carbon Dot Nanoribbons for Optoelectronic Memristor Devices and Neuromorphic Computing.

Carbon dots (CDs) are widely utilized in sensing, energy storage, and catalysis due to their excellent optical, electrical and semiconducting properties. However, attempts to optimize their optoelectronic performance through high-order manipulation have met with little success to date. In this study, through efficient packing of individual CDs in two-dimensions, the synthesis of flexible CDs ribbons is demonstrated technically. Electron microscopies and molecular dynamics simulations, show the assembly of CDs into ribbons results from the tripartite balance of π-π attractions, hydrogen bonding, and halogen bonding forces provided by the superficial ligands. The obtained ribbons are flexible and show excellent stability against UV irradiation and heating. CDs ribbons offer outstanding performance as active layer material in transparent flexible memristors, with the developed devices providing excellent data storage, retention capabilities, and fast optoelectronic responses. A memristor device with a thickness of 8 µm shows good data retention capability even after 104 cycles of bending. Furthermore, the device functions effectively as a neuromorphic computing system with integrated storage and computation capabilities, with the response speed of the device being less than 5.5 ns. These properties create an optoelectronic memristor with rapid Chinese character learning capability. This work lays the foundation for wearable artificial intelligence.

2D Organic Materials: Status and Challenges.

In the past few decades, 2D layer materials have gradually become a central focus in materials science owing to their uniquely layered structural qualities and good optoelectronic properties. However, in the development of 2D materials, several disadvantages, such as limited types of materials and the inability to synthesize large-scale materials, severely confine their application. Therefore, further exploration of new materials and preparation methods is necessary to meet technological developmental needs. Organic molecular materials have the advantage of being customizable. Therefore, if organic molecular and 2D materials are combined, the resulting 2D organic materials would have excellent optical and electrical properties. In addition, through this combination, the free design and large-scale synthesis of 2D materials can be realized in principle. Furthermore, 2D organic materials exhibit excellent properties and unique functionalities along with great potential for developing sensors, biomedicine, and electronics. In this review, 2D organic materials are divided into five categories. The preparation methods and material properties of each class of materials are also described in detail. Notably, to comprehensively understand each material's advantages, the latest research applications for each material are presented in detail and summarized. Finally, the future development and application prospects of 2D organic materials are briefly discussed.

Centimetre-scale single crystal α-MoO3: oxygen assisted self-standing growth and low-energy consumption synaptic devices.

High-density storage and neuromorphic devices based on 2D materials are hindered by large-scale growth. Moreover, the lack of a mature mechanism makes it difficult to obtain high-quality single crystals in large-scale 2D materials. In this work, we prepared a centimeter-scale single crystal α-MoO3via an oxygen assisted substrate-free self-standing growth method and mechanism and constructed high-performance synaptic devices based on the centimeter-scale α-MoO3. The oxygen assisted growth mechanism of α-MoO3 was developed from the periodic bond chain theory. The large-scale α-MoO3 is up to 2 cm and exhibits high homogeneity and single crystalline characteristic. Furthermore, with an optimized oxygen partial pressure (18%), the centimeter-scale α-MoO3 makes the as-prepared memristor achieve continuous conductance modulation. Moreover, the trap-controlled electron conducting mechanism of the memristor was demonstrated through I-V curve fitting analysis at various temperatures, in which the high resistance state section demonstrates space-charge-limited conduction (SCLC) mode. Moreover, the as-prepared α-MoO3 memristors exhibit low-energy consumption and well emulate the essential synaptic behaviors including excitatory/inhibitory postsynaptic current, paired-pulse facilitation and long-term plasticity.

Superlow Power Consumption Artificial Synapses Based on WSe2 Quantum Dots Memristor for Neuromorphic Computing.

As the emerging member of zero-dimension transition metal dichalcogenide, WSe2 quantum dots (QDs) have been applied to memristors and exhibited better resistance switching characteristics and miniaturization size. However, low power consumption and high reliability are still challenges for WSe2 QDs-based memristors as synaptic devices. Here, we demonstrate a high-performance, superlow power consumption memristor device with the structure of Ag/WSe2 QDs/La0.3Sr0.7MnO3/SrTiO3. The device displays excellent resistive switching memory behavior with a R OFF/R ON ratio of ~5 × 103, power consumption per switching as low as 0.16 nW, very low set, and reset voltage of ~0.52 V and~ -0.19 V with excellent cycling stability, good reproducibility, and decent data retention capability. The superlow power consumption characteristic of the device is further proved by the method of density functional theory calculation. In addition, the influence of pulse amplitude, duration, and interval was studied to gradually modulating the conductance of the device. The memristor has also been demonstrated to simulate different functions of artificial synapses, such as excitatory postsynaptic current, spike timing-dependent plasticity, long-term potentiation, long-term depression, and paired-pulse facilitation. Importantly, digit recognition ability based on the WSe2 QDs device is evaluated through a three-layer artificial neural network, and the digit recognition accuracy after 40 times of training can reach up to 94.05%. This study paves a new way for the development of memristor devices with advanced significance for future low power neuromorphic computing.

Integrated Navigation Algorithm Based on Multiple Fading Factors Kalman Filter.

An integrated navigation algorithm based on a multiple fading factors Kalman filter (MFKF) is proposed to solve the problems that the Kalman filtering (KF) algorithm easily brings about diffusion when the model becomes a mismatched or noisy, and the MFKF accuracy is reduced when the fading factor is overused. Based on the innovation covariance theory, the algorithm designs an improved basis for judging filtering anomalies and makes the timing of the introduction of the fading factor more reasonable by switching the filtering state. Different from the traditional basis of filter abnormality judgment, the improved judgment basis adopts a recursive way to continuously update the estimated value of the innovation covariance to improve the estimation accuracy of the innovation covariance, and an empirical reserve factor for the judgment basis is introduced to adapt to practical engineering applications. By establishing an inertial navigation system (INS)/global navigation satellite system (GNSS) integrated navigation model, the results show that the average positioning accuracy of the proposed algorithm is improved by 26.52% and 7.48%, respectively, compared with the KF and MFKF, and shows better robustness and self-adaptability.