Particle filter is commonly used in various indoor positioning schemes, but sample impoverishment and weight degradation generally exist in particle filter. To solve this problem, this article uses the firefly algorithm to optimize particle filter. Using the global optimal value in the particle swarm to guide the remaining particles to update their positions, the particles tend to move to the high likelihood area, which can more accurately describe the true state of the target observation. At the same time, to avoid particles falling into local optimum and oscillating repeatedly at the extreme point, the mathematical model of the firefly algorithm is reconstructed in this article, which is named as adaptive optimization firefly algorithm (AOFA). The experimental results show that the positioning scheme can provide the positioning accuracy with an average error of less than 0.5 m. Compared with the traditional particle filter, the positioning accuracy is improved by 120%.

In this paper, based on the study of Linear Frequency Modulated Continuous Wave (LFMCW) radar, a ranging algorithm for the high-precision liquid level measurement in industry is proposed. A new algorithm is developed based on the MacLeod algorithm. The frequency offset δ ₀ is estimated, and then (δ ₀ -q, δ ₀ +q) is used as the iterative estimation initial iteration interval to determine the frequency offset δ ₀ more precisely. The improved algorithm is the Golden-MacLeod algorithm, referred to as the GM algorithm. Simulation results show that the GM can effectively reduce the estimation error and maintain the high frequency estimation accuracy at signal-to-noise ratio (SNR) up to -10 dB, which is better than the MacLeod algorithm. In addition, the estimation accuracy, stability and noise immunity are better than other algorithms in the variation interval of SNR greater than -5 dB. Liquid level measurement experiment is tested by TI’s IWR 6843ISK and DCA1000EVM. The measurement results also show that the measurement accuracy of the GM can meet the requirements of industrial applications. This also proves the reliability of the GM in practical applications.

This paper proposed an autonomous multi-frame fusion method of millimeter wave (mmWave) radar point cloud suitable for low crowd density indoor scenes to overcome the problem of sparse target points in the application of frequency modulated continuous wave (FMCW) radar in indoor 4D point cloud imaging. Without other sensors, in the static or translational state of the radar, the static and dynamic target points in the radar field of vision are distinguished through multiple velocity iterations, and then the static target points are used to estimate the velocity of the radar itself. By calculating the displacement of the radar within a frame time, we carry out velocity filtering on the point cloud to remove the target points with large differences. Finally, the radar point cloud data of each frame is converted to the same geographic coordinate system to achieve 4D point cloud multi-frame fusion. The experimental results show that the presented method can accurately estimate the velocity of the radar and correct the coordinates of each frame point cloud. According to the imaging results, the proposed algorithm can greatly increase the imaging density of point cloud without defocusing, which improves the accuracy and readability of point cloud image with the imaging ability of static and moving targets.

A continuous HAR method based on the multi-object recognition method (MOR), the constructed lightweight network (LNet), and the proposed one-dimensional bounding loss function (ODBL), the MOR-LNOD (MOR LNet ODBL) method is proposed. The method is validated using continuous action sequences involving nine participants and eight different actions. We interpret each action in the sequence as a single target and utilize a multi-object detection method for accurate single-action region selection, followed by recognition and classification. Based on the results of the study, the MOR-LNOD method is 96.5% accurate on average, which is an improvement of about 20% compared to previous methods based on recurrent neural networks (RNNs). Compared to the Resnet 50 and Mobile net used in traditional Faster R-CNN, the proposed network architecture has reduced the parameters by 10 times and 2 times, respectively. As compared to SOTA (state-of the art) on the publicly available dataset, MOR-LNOD not only reduces the requirement of input data but also has a higher average accuracy than SOTA.

Due to the large number of targets and the close detection distance in indoor environments, a significant amount of ghost target points often appears in the imaging results of mmWave radar point clouds. These ghost points can distort the point cloud images and greatly reduce their readability. Nowadays, ghost elimination methods may affect the imaging accuracy of other targets, and the algorithm is complex, which is not suitable for application in narrow indoor environment. To mitigate interference caused by ghost image target points on already pointed cloud images, this paper proposes a ghost image removal method called Double-Parameter Elimination of Ghost Images (DPEGI) based on point cloud clustering. This method determines the occlusion relationship between each point cloud cluster by setting two key parameters. According to the relative scattering coefficient of the point cloud cluster, the ghost can be effectively removed while retaining the non-line-of-sight (NLOS) targets. Extensive verification using indoor measured data demonstrates that the proposed ghost elimination algorithm is capable of handling common indoor scenes while preserving over 90% of details with high accuracy, robustness, efficiency, and real-time processing capability.

We propose a method for recognizing continuous indoor daily human activities among continuous indoor IoT (Internet of Things) smart environments using millimeter wave radar. Focusing on the following problems: (1) Transition errors due to random transitions between actions during continuous action recognition. (2) The time-consuming and labor-intensive factors on radar data acquisition for continuous human actions make it difficult to consider all action sequences, and the network suffers from catastrophic degradation of the recognition performance when faced with completely new human action sequences. A Bounding Adaptive Optimization Algorithm based on Interlacing Error (BAOA) and a Generative Adversarial Network based on Adaptive Feature Selection Generator and Temporal Patch Discriminator (AFSG-TPD GAN) are proposed. BAOA is used to accurately segment the existing action sequences to obtain a single action dataset of random duration, synthesize the data used to train the AFSG-TPD GAN, generate new action sequences, and train the recognition network to improve generalization performance. After the comparison test, BAOA increases the average accuracy by 3.91% compared to the State-Of-The-Art (SOTA) method. Meanwhile, the network trained with the data generated by the AFSG-TPD GAN overcomes the problem of catastrophic degradation of the recognition performance when confronted with brand new human action sequences in real-world tests, and the average accuracy is improved from 65.01% to 94.85%.