A systematic privacy-preserving framework is proposed in this paper to protect SMS data, using homomorphic encryption with trust boundaries tailored for different SMS applications. We investigated the practicality of the proposed HE framework by measuring its computational performance on two key metrics, summation and variance. These metrics are commonly applied in situations involving billing, usage forecasting, and relevant tasks. In order to secure a 128-bit security level, the security parameters were set appropriately. The performance of calculating the previously mentioned metrics demonstrated 58235 ms for summation and 127423 ms for variance, based on a sample size of 100 households. The proposed HE framework's ability to maintain customer privacy within SMS is corroborated by these results, even under varying trust boundary conditions. Data privacy is preserved, and the computational overhead is justifiable from a cost-benefit standpoint.
Mobile machines, thanks to indoor positioning, can execute tasks (semi-)automatically, like tracking an operator. However, the efficacy and safety of these applications are determined by the trustworthiness of the calculated operator's location. Thus, the process of measuring the accuracy of positioning at runtime is of paramount importance for the application's practical use in industrial settings. This study presents a method that yields an estimation of the current positioning error for each user stride. We use Ultra-Wideband (UWB) location data to formulate a virtual stride vector for this undertaking. Stride vectors, sourced from a foot-mounted Inertial Measurement Unit (IMU), are subsequently used to compare the virtual vectors. Considering these independent measurements, we determine the present accuracy of the UWB data. Positioning errors are alleviated by implementing a loosely coupled filtering system for both vector types. Utilizing three different settings for evaluation, we found our method consistently improved positioning accuracy, especially in challenging environments with limited line of sight and inadequate UWB infrastructure. Subsequently, we illustrate the methods to neutralize simulated spoofing attacks affecting UWB position determination. The process of evaluating positioning quality entails comparing user strides reconstructed from ultra-wideband and inertial measurement unit readings in real time. Our method, which avoids the need for adjusting parameters specific to a given situation or environment, presents a promising avenue for identifying both known and unknown positioning error states.
In Software-Defined Wireless Sensor Networks (SDWSNs), Low-Rate Denial of Service (LDoS) attacks are currently among the most pressing security concerns. Selleck KRpep-2d Network resources are consumed by a flood of low-impact requests, making this kind of attack challenging to discern. The efficiency of LDoS attack detection has been enhanced through a method employing the characteristics of small signals. LDoS attack-generated small, non-smooth signals are scrutinized using time-frequency analysis via Hilbert-Huang Transform (HHT). This paper details the removal of redundant and similar Intrinsic Mode Functions (IMFs) from standard HHT procedures to optimize computational resources and prevent modal interference. Dataflow features, originally one-dimensional, were transformed into two-dimensional temporal-spectral characteristics via the compressed Hilbert-Huang Transform (HHT) and subsequently fed into a Convolutional Neural Network (CNN) to identify LDoS attacks. Using the NS-3 simulator, the detection performance of the method was assessed by carrying out simulations of different LDoS attack types. The experimental results affirm the method's ability to detect complex and diverse LDoS attacks with an accuracy of 998%.
Deep neural networks (DNNs) are vulnerable to backdoor attacks, a technique that triggers misclassifications. An adversary seeking to activate a backdoor attack introduces an image bearing a specific pattern (the adversarial marker) into the DNN model (specifically, the backdoor model). A photograph of the physical input object is usually required to establish the adversary's mark. The conventional backdoor attack method's success rate is unstable, with size and location variations influenced by the shooting environment. Our earlier work introduced a technique for creating an adversarial signal designed to activate backdoor attacks via fault injection on the MIPI, the image sensor's communication interface. Employing actual fault injection, our proposed image tampering model produces adversarial marks, resulting in a structured adversarial marker pattern. Training of the backdoor model was subsequently performed utilizing data images containing malicious elements; these images were created by the proposed simulation model. Using a backdoor model trained on a dataset with 5% poisoned data, our experiment investigated backdoor attacks. Microbiota-independent effects Normal operation maintained a 91% clean data accuracy; however, fault injection led to an 83% attack success rate.
For carrying out dynamic mechanical impact tests on civil engineering structures, shock tubes are employed. The predominant method used in current shock tubes involves an explosion utilizing an aggregated charge to achieve shock waves. Despite the critical importance of studying the overpressure field in shock tubes with multi-point initiation, limited resources and effort have been applied. The pressure surge characteristics in shock tubes, triggered by single-point, simultaneous multi-point, and sequential multi-point ignition, are explored in this paper through a combination of experimental observations and numerical simulations. The shock tube's blast flow field is accurately simulated by the computational model and method, as corroborated by the remarkable concordance between the numerical results and experimental data. Under identical charge mass conditions, the peak overpressure recorded at the shock tube's outlet is lower for multiple simultaneous initiation points as opposed to a single initiation point. Despite the focusing of shock waves on the wall, the extreme pressure exerted upon the explosion chamber's wall close to the explosion remains unchanged. By utilizing a six-point delayed initiation, the maximum overpressure exerted on the explosion chamber's wall is significantly reduced. Under the condition of an explosion interval less than 10 milliseconds, the peak overpressure at the nozzle's exit demonstrates a linear decline in accordance with the interval's duration. For interval times exceeding 10 milliseconds, the overpressure peak is unaffected.
The complex and hazardous working conditions of human forest operators have made automated forest machinery a critical necessity, effectively mitigating the labor shortage problem. Employing low-resolution LiDAR sensors, this study proposes a novel and robust simultaneous localization and mapping (SLAM) methodology for tree mapping within forestry environments. Oil biosynthesis Our scan registration and pose correction process, reliant on tree detection, operates exclusively with low-resolution LiDAR sensors (16Ch, 32Ch) or narrow field of view Solid State LiDARs, thereby dispensing with extra sensory inputs like GPS or IMU. Three datasets—two internal and one public—were used to evaluate our approach, showing an improvement in navigation accuracy, scan alignment, tree localization, and tree girth estimation compared to the current state-of-the-art in forestry machine automation. Employing detected trees, the proposed method achieves robust scan registration, surpassing the performance of generic feature-based algorithms such as Fast Point Feature Histogram by over 3 meters in RMSE, specifically with the 16-channel LiDAR. In the case of Solid-State LiDAR, a similar RMSE of 37 meters is obtained by the algorithm. The adaptive pre-processing, coupled with a heuristic tree detection approach, increased the number of identified trees by 13% compared to the existing pre-processing method using fixed radius search parameters. By employing automated estimation of tree trunk diameters on local and complete trajectory maps, we observe a mean absolute error of 43 cm; the root mean squared error is 65 cm.
The popularity of fitness yoga has significantly impacted the national fitness and sportive physical therapy landscape. Microsoft Kinect, a depth-sensing apparatus, and various other applications for yoga are in widespread use to assess and direct performance, however, practical application is limited by their expense and complexity. Employing spatial-temporal self-attention mechanisms within graph convolutional networks (STSAE-GCNs), we aim to resolve these problems by examining RGB yoga video data captured by cameras or smartphones. Within the STSAE-GCN architecture, a spatial-temporal self-attention module (STSAM) is constructed, significantly boosting the spatial-temporal representational capacity of the model and thereby enhancing its overall performance. The STSAM's plug-and-play characteristics facilitate its integration into existing skeleton-based action recognition systems, thereby improving their overall performance. We constructed the Yoga10 dataset, comprising 960 video clips of fitness yoga actions, categorized across 10 action classes, to evaluate the effectiveness of our proposed model in recognizing these actions. The Yoga10 dataset reveals a 93.83% recognition accuracy for this model, an improvement over the leading techniques, emphasizing its enhanced capacity to identify fitness yoga actions and facilitate autonomous student learning.
Precisely quantifying water quality is essential for effective monitoring of aquatic environments and responsible water resource management, and has become integral to ecological recovery and sustainable progress. Still, the marked spatial disparities in water quality parameters make it difficult to ascertain highly accurate spatial patterns. From the perspective of chemical oxygen demand, this study develops a novel method for creating highly accurate chemical oxygen demand fields, specifically in Poyang Lake. An optimal virtual sensor network, specifically designed for Poyang Lake, was initially established, taking into account variations in water levels and monitoring sites.