Search

Recently, with the rapid development of automatic driving, more and more researchers begin to explore the lightweight of image semantic segmentation network and apply it to road traffic scenes. However, the existing semantic segmentation networks are usually difficult to deploy in edge devices with limited hardware resources due to the large number of parameters. Aiming at solving this problem, a rapid and ultra-lightweight dual attention lightweight network (DALNet) composed of channel attention bottleneck backbone (CABb) network and spatial attention decoder (SAD) module is proposed in this paper, which has outstanding performance in extracting the context semantic information and spatial information of the image. The CABb network is mainly composed of channel attention bottleneck (CABt) module. Split strategy is employed in CABt to separate feature channels and process multi-scale feature maps in parallel. And channel attention mechanism is introduced for channel fusion and multi-scale semantic information extraction. The spatial attention mechanism is adopted in SAD module to guide the decoder to upsample the feature maps using bilinear interpolation and recover the edge information and detail information of segmentation target. Experimental results show that the proposed DALNet has only 0.48 million parameters and achieves 74.1% and 70.1% mean intersection over union (mIoU) in the popular urban traffic datasets of Cityscapes and CamVid. With the resolution of 512×1024, DALNet achieves 74 frame/s inference speed on a GTX 1080Ti card, which meets the speed requirements of real-time semantic segmentation adequately.

Identifying the underlying surface at varying speeds during robotic locomotion is important for safe and efficient robot navigation. This work aims to enhance the perceptual abilities of biped robots by enabling the force sensor affixed beneath each foot to recognize multiple indoor surfaces while navigating at varying speeds. An accurate yet cost-efficient surface-identification system for the robot is proposed by combining a real-time multiobject support vector machine (SVM) with an efficient time-domain feature. In this context, four promising hand-crafted time-domain features are investigated, among which the root mean square (RMS) feature is proven to outperform the other three features. A mean average precision (mAP) of 95.99% and 98.16% in tenfold SVM cross-validation can be achieved by applying RMS at two different walking speeds, respectively. The high classification accuracy is achieved with high computing efficiency and thus enables system deployment on low-cost platforms, such as Arduino or Jetson Nano, which makes our approach suitable for a wide range of applications across varying walking speeds. [ABSTRACT FROM AUTHOR]