Accurate and up-to-date 3D maps, often represented as point clouds, are crucial for autonomous vehicles. Crowd-sourcing has emerged as a low-cost and scalable approach for collecting mapping data utilizing widely available dashcams and other sensing devices. However, it is still a non-trivial task to utilize crowdsourced data, such as dashcam images and video, to efficiently create or update high-quality point clouds using technologies like Structure from Motion (SfM). This study assesses and compares different image matching options available in open-source SfM software, analyzing their applicability and limitations for mapping urban scenes in different practical scenarios. Furthermore, the study analyzes the impact of various camera setups (i.e., the number of cameras and their placement) and weather conditions on the quality of the generated 3D point clouds in terms of completeness and accuracy. Based on these analyses, our study provides guidelines for creating more accurate point clouds.

Up-to-date and accurate road maps with traffic rules are crucial for traffic safety and efficiency. However, detecting changes in traffic rules, particularly in road signs in an efficient manner still remains an open challenge. This paper proposes a method to detect and classify seven types of changes, relying on observable traffic flow changes like average vehicle speed. Our method employs a deep learning-based binary relevance approach, treating each type of change as a separate binary classification task. Input data comprise information from citywide induction loops, detailing congestion levels and average speed for each road, which vary with time and location. Our model outputs change type probabilities for 2.5 km2city regions for every 10 minutes. Unlike GPS traces of buses and taxis, which provide merely a partial view of the traffic, induction loop data offers a comprehensive, cost-effective view of traffic. However, there are three key challenges of utilizing induction loop data for change detection and classification, including sparse deployment of induction loops, high dimensionality of input data, and severe class imbalance. To address the first two challenges, we introduce novel strategies comprising dimensionality reduction, multi-dimensional sliding windows, neural network architectural choices such as residual connections and stateful recurrency, and data augmentation. Regarding the class imbalance, we apply sample weighing in the loss function. These novel design choices result in effective solutions with F1-scores exceeding 80% in experiments with simulated and real-world traffic data.

Autonomous driving requires 3-D maps that provide accurate and up-to-date information about semantic landmarks. Since cameras present wider availability and lower cost compared with laser scanners, vision-based mapping solutions, especially, the ones using crowdsourced visual data, have attracted much attention from academia and industry. However, previous works have mainly focused on creating 3-D point clouds, leaving automatic change detection as open issue. We propose a pipeline for initiating and updating 3-D maps with dashcam videos, with a focus on automatic change detection based on comparison of metadata (e.g., the types and locations of traffic signs). To improve the performance of metadata generation, which depends on the accuracy of 3-D object detection and localization, we introduce a novel deep learning-based pixelwise 3-D localization algorithm. The algorithm, trained directly with Structure from Motion (SfM) point cloud data, accurately locates objects in 3-D space by estimating not only depth from monocular images but also lateral and height distances. In addition, we also propose a point clustering and thresholding algorithm to improve the robustness of the system to errors. We have performed experiments with different types of cameras, lighting, and weather conditions. The changes were detected with an average accuracy above 90%. The errors in the campus area were mainly due to traffic signs seen from a far distance to the vehicle and intended for pedestrians and cyclists only. We also conducted cause analysis of the detection and localization errors to measure the impact from the performance of the background technology in use.

Recent advances in deep learning have granted unrivaled performance to sensor-based human activity recognition (HAR) . However, in a real-world scenario, the HAR solution is subject to diverse changes over time such as the need to learn new activity classes or variations in the data distribution of the already-included activities. To solve these issues, previous studies have tried to apply directly the continual learning methods borrowed from the computer vision domain, where it is vastly explored. Unfortunately, these methods either lead to surprisingly poor results or demand copious amounts of computational resources, which is infeasible for the low-cost resource-constrained devices utilized in HAR. In this paper, we provide a resource-efficient and high-performance continual learning solution for HAR. It consists of an expandable neural network trained with a replay-based method that utilizes a highly-compressed replay memory whose samples are selected to maximize data variability. Experiments with four open datasets, which were conducted on two distinct microcontrollers, show that our method is capable of achieving substantial accuracy improvements over baselines in continual learning such as Gradient Episodic Memory, while utilizing only one-third of the memory and being up to 3× faster.

Deep learning has permitted unprecedented performance in sensor-based human activity recognition (HAR). However, deep learning models often present high computational overheads, which poses challenges to their implementation on resource-constraint devices such as microcontrollers. Usually, the computational overhead increases with the input size. One way to reduce the input size is by constraining the number of sensor channels. We refer to sensor channel as a specific data modality (e.g. accelerometer) placed on a specific body location (e.g. chest). Identifying and removing irrelevant and redundant sensor channels is feasible via exhaustive search only in cases where few candidates exist. In this paper, we propose a smarter and more efficient way to optimize the sensor channel selection during the training of deep neural networks for HAR. Firstly, we propose a light-weight deep neural network architecture that learns to minimize the use of redundant and irrelevant information in the classification task, while achieving high performance. Secondly, we propose a sensor channel selection algorithm that utilizes the knowledge learned by the neural network to rank the sensor channels by their contribution to the classification task. The neural network is then trimmed by removing the sensor channels with the least contribution from the input and pruning the corresponding weights involved in processing them. The pipeline that consists of the above two steps iterates until the optimal set of sensor channels has been found to balance the trade-off between resource consumption and classification performance. Compared with other selection methods in the literature, experiments on 5 public datasets showed that our proposal achieved significantly higher F1-scores at the same time as utilizing from 76% to 93% less memory, with up to 75% faster inference time and as far as 76% lower energy consumption.