In this study, a millimeter-wave (MMW) radar and an onboard camera are used to develop a sensor fusion algorithm for a forward collision warning system. This study proposed integrating an MMW radar and camera to compensate for the deficiencies caused by relying on a single sensor and to improve frontal object detection rates. Density-based spatial clustering of applications with noise and particle filter algorithms are used in the radar-based object detection system to remove non-object noise and track the target object. Meanwhile, the two-stage vision recognition system can detect and recognize the objects in front of a vehicle. The detected objects include pedestrians, motorcycles, and cars. The spatial alignment uses a radial basis function neural network to learn the conversion relationship between the distance information of the MMW radar and the coordinate information in the image. Then a neural network is utilized for object matching. The sensor with a higher confidence index is selected as the system output. Finally, three kinds of scenario conditions (daytime, nighttime, and rainy-day) were designed to test the performance of the proposed method. The detection rates and the false alarm rates of proposed system were approximately 90.5% and 0.6%, respectively.

The Internet of things (IoT) enables mobile devices to connect and exchange information with others over the Internet with a lot of applications in consumer, commercial, and industrial products. With the rapid development of machine learning, IoT with image recognition capability is a new research area to assist mobile devices with processing image information. In this research, we propose the rotation-invariant multi-scale convolutional neural network (RIMS-CNN) to recognize rotated objects, which are commonly seen in real situations. Based on the dihedral group D4 transformations, the RIMS-CNN equips a CNN with multiple rotated tensors and its processing network. Furthermore, multi-scale features and shared weights are employed in the RIMS-CNN to increase performance. Compared with the data augmentation approach of using rotated images at random angles for training, our proposed method can learn inherent convolution kernels for rotational features. Experiments were conducted on the benchmark datasets: MNIST, FASHION-MNIST, CIFAR-10, and CIFAR-100. Significant improvements over the other models were achieved to show that rotational invariance could be learned.