When, Where and How Does it Fail? A Spatial-Temporal Visual Analytics Approach for Interpretable Object Detection in Autonomous Driving.

Arguably the most representative application of artificial intelligence, autonomous driving systems usually rely on computer vision techniques to detect the situations of the external environment. Object detection underpins the ability of scene understanding in such systems. However, existing object detection algorithms often behave as a black box, so when a model fails, no information is available on When, Where and How the failure happened. In this paper, we propose a visual analytics approach to help model developers interpret the model failures. The system includes the micro- and macro-interpreting modules to address the interpretability problem of object detection in autonomous driving. The micro-interpreting module extracts and visualizes the features of a convolutional neural network (CNN) algorithm with density maps, while the macro-interpreting module provides spatial-temporal information of an autonomous driving vehicle and its environment. With the situation awareness of the spatial, temporal and neural network information, our system facilitates the understanding of the results of object detection algorithms, and helps the model developers better understand, tune and develop the models. We use real-world autonomous driving data to perform case studies by involving domain experts in computer vision and autonomous driving to evaluate our system. The results from our interviews with them show the effectiveness of our approach.

Visual Evaluation for Autonomous Driving.

Autonomous driving technologies often use state-of-the-art artificial intelligence algorithms to understand the relationship between the vehicle and the external environment, to predict the changes of the environment, and then to plan and control the behaviors of the vehicle accordingly. The complexity of such technologies makes it challenging to evaluate the performance of autonomous driving systems and to find ways to improve them. The current approaches to evaluating such autonomous driving systems largely use a single score to indicate the overall performance of a system, but domain experts have difficulties in understanding how individual components or algorithms in an autonomous driving system may contribute to the score. To address this problem, we collaborate with domain experts on autonomous driving algorithms, and propose a visual evaluation method for autonomous driving. Our method considers the data generated in all components during the whole process of autonomous driving, including perception results, planning routes, prediction of obstacles, various controlling parameters, and evaluation of comfort. We develop a visual analytics workflow to integrate an evaluation mathematical model with adjustable parameters, support the evaluation of the system from the level of the overall performance to the level of detailed measures of individual components, and to show both evaluation scores and their contributing factors. Our implemented visual analytics system provides an overview evaluation score at the beginning and shows the animation of the dynamic change of the scores at each period. Experts can interactively explore the specific component at different time periods and identify related factors. With our method, domain experts not only learn about the performance of an autonomous driving system, but also identify and access the problematic parts of each component. Our visual evaluation system can be applied to the autonomous driving simulation system and used for various evaluation cases. The results of using our system in some simulation cases and the feedback from involved domain experts confirm the usefulness and efficiency of our method in helping people gain in-depth insight into autonomous driving systems.

Gradient structured micro/nanofibrous sponges with superior compressibility and stretchability for broadband sound absorption.

Ultrafine fibrous porous materials obtained by electrospinning technology have broad application prospects in the field of noise reduction. However, the two-dimensional fibrous membranes faced low thickness and dense structure, resulting in a single internal structure and narrow sound absorption band. Here, we report a simple and robust strategy to prepare gradient structured fiber sponges with superelasticity and stretchability by combining humidity-assisted multi-step electrospinning and a unique physical/chemical dual cross-linking method. The prepared gradient structured fibrous sponge has a maximum tensile strength of 169 kPa and can lift a weight 10,000 times its weight without breaking. Besides, the material can still maintain a stable structure after 500 compression cycles at 60% strain. Meantime, the material has lightweight properties (density of 13.8 mg cm-3) and hydrophobicity (water contact angle of 152°). More importantly, the gradient change of porosity and pore diameter in the Z direction endowed the fibrous sponge material with high-efficiency absorption of broadband sound waves (with a noise reduction coefficient up to 0.53). The design of this gradient structured fiber sponge opens a new way for the development of ideal sound-absorbing materials.

Exploiting Kalman Filtering Non-linear Exponential Fitting to Promote the Energy Resolution of 137Cs and 60Co Gamma Ray Spectra.

For Cs and Co gamma ray spectra, gamma ray energy is proportional to the amplitude of the pulse signal, and energy resolution can be improved by pulse signal processing with mathematical algorithms. Influenced by system measurement noise and baseline fluctuation, the pulse amplitude is difficult to calculate accurately. A method that combines the Kalman filter baseline estimation with the non-linear exponential fitting has been used. By this method, the pulse signal is divided into two parts: one is the raising edge before the pulse peak, and another is after the pulse peak. The pulse amplitude equals the difference between the pulse starting height and the pulse peak height. The pulse starting height is obtained by Kalman filter baseline estimation on the rising edge of the pulse starting point. The pulse peak height is calculated by nonlinear exponential fitting on the falling edge of the pulse highest point. When the sampling rate is 100 MHz, the pulse signals obtained from a Cd(Zn)Te detector are analyzed by this method. Results have shown that the processed pulses have a more distinguishable amplitude distribution; energy resolution for the Cs spectrum is approximately 2.97% at 662 keV (~19.66 keV FWHM), and for the Co spectrum it is 2.61% at 1,332 keV (~34.76 keV FWHM).