Convolutional neural networks (CNNs) have been widely deployed in the fields of computer vision and pattern recognition because of their high accuracy. However, large convolution operations are computing intensive and often require a powerful computing platform such as a graphics processing unit. This makes it difficult to apply CNNs to portable devices. The state-of-the-art CNNs, such as MobileNetV2 and Xception, adopt depthwise separable convolution to replace the standard convolution for embedded platforms, which significantly reduces operations and parameters with only limited loss in accuracy. This highly structured model is very suitable for field-programmable gate array (FPGA) implementation. In this brief, a scalable high performance depthwise separable convolution optimized CNN accelerator is proposed. The accelerator can be fit into an FPGA of different sizes, provided the balancing between hardware resources and processing speed. As an example, MobileNetV2 is implemented on Arria 10 SoC FPGA, and the results show this accelerator can classify each picture from ImageNet in 3.75 ms, which is about 266.6 frames per second. The FPGA design achieves 20× speedup if compared to CPU.

Respiration rate (RR) and heartbeat rate (HR) are important physiological parameters for a person. Impulse radio ultra-wideband (IR-UWB) is a promising technology for non-contact sensing and monitoring. This brief presents a new method based on autocorrelation to measure the RR and HR using IR-UWB radar. The correlation coefficient waveform contains the vital sign signals, overcoming the effect of noise and clutter. Applying fast Fourier transform, the respiration frequency can be acquired easily. A clever method also based on autocorrelation is proposed to locate the subject. The receive signal matrix is divided into a set of bins in the direction of fast time. By removing one block from the matrix each time and re-applying the autocorrelation, the removed block resulting the smallest correlations is corresponding to the location of a subject. Moreover, variational mode decomposition algorithm is adopted to successfully separate the respiration and heartbeat signals. Experiments are carried out using a PulsOn410 UWB radar. The results show that the proposed low-complexity algorithm has high accuracy.

Compressed sensing allows to reconstruct sparse signals sampled at sub-Nyquist rates. However, reconstruction of the original signal requires high computational effort, even for problems of moderate size. Especially for applications with real-time requirements, software realizations are not fast enough. We therefore present generic high-speed FPGA implementations of two fast reconstruction algorithms: orthogonal matching pursuit (OMP) and approximate message passing (AMP). Our implementations also support less sparse signals, which makes them suitable for, e.g., image reconstruction. The two implementations are optimized for highly parallel processing on FPGAs and have similar hardware structures, which allows comparisons in terms of resource usage and performance.

LiDAR depth completion is a task that predicts depth values for every pixel on the corresponding camera frame, although only sparse LiDAR points are available. Most of the existing state-of-the-art solutions are based on deep neural networks, which need a large amount of data and heavy computations for training the models. In this letter, a novel non-learning depth completion method is proposed by exploiting the local surface geometry that is enhanced by an outlier removal algorithm. The proposed surface geometry model is inspired by the observation that most pixels with unknown depth have a nearby LiDAR point. Therefore, it is assumed those pixels share the same surface with the nearest LiDAR point, and their respective depth can be estimated as the nearest LiDAR depth value plus a residual error. The residual error is calculated by using a derived equation with several physical parameters as input, including the known camera intrinsic parameters, estimated normal vector, and offset distance on the image plane. The proposed method is further enhanced by an outlier removal algorithm that is designed to remove incorrectly mapped LiDAR points from occluded regions. On KITTI dataset, the proposed solution achieves the best error performance among all existing non-learning methods and is comparable to the best self-supervised learning method and some supervised learning methods. Moreover, since outlier points from occluded regions is a commonly existing problem, the proposed outlier removal algorithm is a general preprocessing step that is applicable to many robotic systems with both camera and LiDAR sensors. The code has been published at https://github.com/placeforyiming/RAL_Non-Learning_DepthCompletion.

In automated driving systems (ADS) and advanced driver-assistance systems (ADAS), an efficient road segmentation is necessary to perceive the drivable region and build an occupancy map for path planning. The existing algorithms implement gigantic convolutional neural networks (CNNs) that are computationally expensive and time consuming. In this paper, we introduced distributed LSTM, a neural network widely used in audio and video processing, to process rows and columns in images and feature maps. We then propose a new network combining the convolutional and distributed LSTM layers to solve the road segmentation problem. In the end, the network is trained and tested in KITTI road benchmark. The result shows that the combined structure enhances the feature extraction and processing but takes less processing time than pure CNN structure.