System Resource Constraints Research Articles

Optimizing embedded AI systems for autonomous driving: Challenges and solutions using bayesian networks

Abstract. Embedded Artificial Intelligence (AI) systems are important components of autonomous vehicles. However, incorporating AI into autonomous vehicles is technically complex, due to the constraints of computation, real-time processing of data, uncertainty handling, and hardware limitations. Bayesian Networks (BNs) are a promising method that allows probabilistic modelling in adaptive learning and environment perception. Here we report on an overview of the application of BNs on autonomous driving, with an emphasis on how BNs can be optimized for embedded system resource constraints, including both computational and energy. Various optimization techniques are discussed, such as model pruning, approximation, acceleration using hardware accelerators, such as Field-Programmable Gate Arrays (FPGA) and Application Specific Integrated Circuit (ASICs), and advanced cooling and power management to ensure AI reliability under high computational load. By reviewing these approaches, we aim to contribute to the development of more robust and green AI systems for autonomous driving.

Economic Assessment in Resource-Constrained Systems: Individual-Level Simulation Model in Wet Age-Related Macular Degeneration and Diabetic Macular Oedema.

Cost-effectiveness analyses typically ignore healthcare system resource constraints. Ophthalmology is affected by resource constraints because of increasing disease prevalence and the use of resource-intensive treatments. This study evaluated the impact of resource constraints on the cost-effectiveness of faricimab 6mg, compared with aflibercept 2mg and ranibizumab biosimilar 0.5mg, for treating wet age-related macular degeneration (wAMD) or diabetic macular oedema (DMO) over a 5-year horizon. A microsimulation model estimated the impact of resource constraints on patients visits, delays, costs and quality-adjusted life-year (QALY) losses due to treatment delays at a typical UK National Health Service eye hospital treating 1500 patients with wAMD and 500 patients with DMO. Patient characteristics, treatment regimens and treatment intervals were informed using published literature and expert opinion. Resource constraint was represented by limiting the number of available intravitreal injection appointments per week, with growing demand caused by rising disease prevalence. The model compared outcomes across three scenarios; each scenario involved treating all patients with one of the three treatments. Over 5years, in a resource-constrained hospital, compared with aflibercept, faricimab use resulted in the avoidance of 12,596 delays, saved GBP/£15,108,609 in cost and avoided the loss of 60.06 QALYs. Compared with ranibizumab biosimilar, faricimab use resulted in the avoidance of 18,910 delays, incurred £2,069,088 extra cost and avoided the loss of 105.70 QALYs, resulting in an incremental cost-effectiveness ratio of £19,574/QALY. Accounting for resource constraints in health economic evaluation is crucial. Emerging therapies that are more durable and require less frequent clinic visits can reduce treatment delays, leading to improved QALY outcomes and reduced burden on healthcare systems. Faricimab reduced the number of delayed injections, leading to improved QALY outcomes for patients in a healthcare system with resource constraints. Faricimab is cost-saving when compared with aflibercept and cost-effective when compared with ranibizumab biosimilar.

A pilot to promote early child development within health systems in Mozambique: a qualitative evaluation.

Health systems offer unique opportunities for integrating services to promote early child development (ECD). However, there is limited knowledge about the implementation experiences of using health services to target nurturing care and ECD, especially in sub-Saharan Africa. We conducted a qualitative implementation evaluation to assess the delivery, acceptability, perceived changes, and barriers and facilitators associated with a pilot strategy that integrated developmental monitoring, nutritional screening, and early learning and nutrition counseling into the existing health facility, and community-based services for young children in rural Mozambique. We completed individual interviews with caregivers (N=36), providers (N=27), and district stakeholders (N=10), and nine facility observational visits at three primary health facilities in October-November 2020. We analyzed data using thematic content analysis. Results supported fidelity to the intended pilot model and acceptability of nurturing care services. Respondents expressed various program benefits, including strengthened health system capacity and improved knowledge, attitudes, and practices regarding nurturing care and ECD. Government leadership and supportive supervision were key facilitators, whereas health system resource constraints were key barriers. We conclude that health systems are promising platforms for supporting ECD and discuss several programmatic recommendations for enhancing service delivery and maximizing potential impacts on nurturing care and ECD outcomes.

Inspection Allocation Optimization with Resource Constraints Based on Modified NSGA-II in Flexible Manufacturing Systems

With the development of smart manufacturing, quality has become an indispensable issue in the manufacturing process. Although there is increasing publication about inspection allocation problems, inspection allocation optimization research considering resource capability is scarce. This paper focuses on the inspection allocation problem with resource constraints in the flexible manufacturing system. Combined with the inspection resource capability model, a bi-objective model is developed to minimize the cost and balance loads of the inspection station. A modified NSGA-II algorithm with adaptive mutation operators is suggested to deal with the proposed model. Finally, a simulation experiment is conducted to test the performance of the modified algorithm and the results demonstrate that modified NSGA-II can obtain acceptable inspection solutions.

Optimal State Risk Scheduling Based on Selective Maintenance Strategy

Because stochastic fault cases (i.e., opportunistic maintenance strategy) of the equipment are not considered in the condition-based maintenance decision of the system, which will deviate from the actual situation, a system condition-based maintenance scheduling model considering opportunistic maintenance strategy is proposed in this paper. To implement the system maintenance strategies, the correlation set is formulated by considering the relationship among different equipment. According to renewal process theory, the availability of the correlation set considering planned maintenance and opportunity maintenance is deduced and the maintenance strategy of the system is realized. Then, to reflect the relationship between equipment maintenance and system operation, a system state scheduling model aiming at minimizing the sum of maintenance risk and failure risk as well as considering system resource constraints is proposed, thus obtaining the optimal maintenance schedule based on system state. Finally, a simple test system and IEEE-RTS79 node system are employed to demonstrate the feasibility and effectiveness of the proposed maintenance model, and it also verified that the proposed model can be integrated into the power system condition-based maintenance theory.

Federated Edge Learning: Design Issues and Challenges

Federated Learning (FL) is a distributed machine learning technique, where each device contributes to the learning model by independently computing the gradient based on its local training data. It has recently become a hot research topic, as it promises several benefits related to data privacy and scalability. However, implementing FL at the network edge is challenging due to system and data heterogeneity and resource constraints. In this article, we examine the existing challenges and trade-offs in Federated Edge Learning (FEEL). The design of FEEL algorithms for resources-efficient learning raises several challenges. These challenges are essentially related to the multidisciplinary nature of the problem. As the data is the key component of the learning, this article advocates a new set of considerations for data characteristics in wireless scheduling algorithms in FEEL. Hence, we propose a general framework for data-aware scheduling as a guideline for future research directions. We also discuss the main axes and requirements for data evaluation and some exploitable techniques and metrics.

Smart Approach for Glioma Segmentation in Magnetic Resonance Imaging using Modified Convolutional Network Architecture (U-NET)

Segmentation of a brain tumor from magnetic resonance multimodal images is a challenging task in the field of medical imaging. The vast diversity in potential target regions, appearance and multifarious intensity threshold levels of various tumor types are few of the major factors that affect segmentation results. An accurate diagnosis and its treatment demand strict delineation of the tumor affected tissues. Herein, we focus on a smart, automated, and robust segmentation approach for brain tumor using a modified 3D U-Net architecture. The pre-operative multimodal 3D-MRI scans of High-Grade Glioma (HGG) and Low-Grade Glioma (LGG) are used as data. Our proposed approach solves the problem of memory and system resource constraints by robustly applying dense network training on image patches of 3D volumes. It improves the border region artifact detection by applying convolutions at an appropriate phase in the proposed neural network. Multi-class imbalance data are handled by using Categorical Cross Entropy (CCE) loss developed by combining the Weighted Cross Entropy (WCE) with Weighted Multi-class Dice Loss (WMDL) functions, which enables the network to perform smart segmentation of the smaller tumorous regions. The proposed approach is tested and evaluated for the challenge datasets of multimodal MRI volumes of tumor patients. Experiments are performed to compute the average dice scores on BraTS-2019 and BraTS-2020 datasets for the whole tumor region.

The Question-prompt list (QPL): Why it is needed in the Indian oncology setting?

BackgroundIn India, caregivers are an integral part of the illness experience, especially in cancer, to the extent that they can become proxy decision‐makers for the patient. Further, owing to acute resource constraints in the Indian healthcare system, it may be difficult for oncologists to assess and elicit questions from each patient/caregiver. Consequently, there is a need to address these unique aspects of oncology care in India to improve patient outcomes and understanding of their illness and treatment. This can be achieved through a Question Prompt List (QPL), a checklist used by care recipients during medical consultations.Recent FindingsThis narrative review will first introduce research on the development and effectiveness of the QPL, and then it will highlight current gaps in oncology care in India and explore how the QPL may aid in closing these gaps.A literature search of the empirical research focused on the development, feasibility and acceptability of the QPL in oncology settings was conducted. The final review included 40 articles pertaining to QPL research. Additionally, psycho‐oncology research in India centered on information needs and experiences was reviewed.Current Indian psycho‐oncology research reports patients' want to be actively involved in their cancer care and a need for more illness information. However, a high demand on physicians' resources and the family caregivers' interference can be barriers to meeting patients' information/communication needs. International research demonstrates that a QPL helps structure and decrease consultation time, improves patient satisfaction with care, and improves the quality of communication during medical encounters.ConclusionQPLs for Indian patients and caregivers may focus on the scope of medical consultations to address patient needs while influencing the course and content of the patient‐caregiver‐physician interactions. Further, it can address the resource constraints in Indian oncology care settings, thus reducing the physician's burden.

DVFS-Based Quality Maximization for Adaptive Applications With Diminishing Return

Application-level approximate computing exploits inherent resilience of adaptive applications, and trades off application output quality for runtime system resources. Existing methods treat computing quality as the number of clock cycles to execute a task, but they overlook the fact that the quality of many real-life applications exhibit the characteristic of diminishing return as the processor continues executing. The diminishing return of the quality is largely due to the features of iterative processing or successive refinement inherent in those applications. Ignoring it leads to large over-estimation in contemporary quality optimization approaches. In this article, we exploit the application adaptability to achieve quality maximization by taking both system resource constraints and diminishing return of the quality into account. We first reveal that the diminishing return of the quality is inherent in several well-known applications, and suggest an exponential model that accurately captures it. Second, we propose a dynamic frequency scaling (DFS) methodology to optimally decide the processor execution cycles for such applications, in order to maximize the output quality under system energy, timing, and temperature constraints. We transform the DFS problem to an iterative pseudo quadratic programming heuristic that can be efficiently solved. Third, we present a wrapping dynamic voltage scaling (wDVS) methodology to achieve further quality improvement, by judiciously adjusting the supply voltage to provide extra frequency scaling space. Compared to state-of-the-art algorithms, our approach produces at least 19.1 percent quality improvement on all evaluated cases, with negligible execution overhead.

Guest Editorial: Trustworthiness in Industrial Internet of Things Systems and Applications

Trustworthiness is the probability that a system will function according to intended behaviors under a set of circumstances as demonstrated by qualities including, but not limited to safety, security, privacy, reliability, real timeliness. Trustworthiness in the industrial Internet of Things (IIoT) systems and applications is crucial to a vital expectation of industrial investors. Preserving the trustworthiness of such a system and network is crucial to void cost, time, and loss of lives. A trustworthy IIoT system considers both the security characteristics and system functionalities (faults, failures) of IoT trustworthiness. Traditional security systems, tools, techniques, and apps are not enough to protect the IIoT platform due to the practical facts in real industrial environments, diverse protocols, constrained upgrade opportunities, mismatch in protocols, and resource constraints in the industrial system. Regarding these concerns, this special section targets to bring up-to-date research results of IIoT systems and applications with the trustworthiness support. This leads to steady operations and high-quality results and improved safety in the IIoT.

Game Theoretic Feedback Control for Reliability Enhancement of EtherCAT-Based Networked Systems

EtherCAT has become one of the leading real-time Ethernet solutions for networked industrial systems, where a reliable communication infrastructure is needed due to highly error-prone environments. However, existing work on EtherCAT mainly focuses on clock synchronization and timeliness improvement. The reliability of EtherCAT-based networked systems has largely been ignored. In this paper, we present a proportional integral derivative (PID)-based feedback control scheme that aims at enhancing reliability of networked systems under timing and system resource constraints. Instead of retransmitting data upon error detection, we use forward error control technique based on inequality of arithmetic and geometric means to achieve the required system reliability at a low deadline miss rate of messages. We further optimize the forward error control technique and design a fast and fair error resilient mechanism by using a cooperative game. In addition to reliability enhancement, our PID-based error control scheme can also improve the stability of a system in terms of deadline miss rate in the presence of burst errors. Simulation results show that the proposed scheme can achieve reliability enhancement of up to 91% compared to benchmarking methods.

Blood pressure control, glycemic control, and dyslipidemia among healthy adults in the Cape Coast metropolis, Ghana

BackgroundThe World Health Organization recommends the implementation of interventions focused on the early detection of clinical risk factors for cardiovascular disease (CVD) as effective strategies for the control of CVD in low resource settings. However, due to health system resource constraints, surveillance capacity for the identification of high-risk populations for non-communicable diseases, including CVD have been inadequate. The purpose of this study was to describe the prevalence of CVD clinical risk factors among healthy adults residing in the Cape Coast metropolis of Ghana. The clinical risk factors assessed included glycemic control, insulin sensitivity, lipid control and blood pressure. MethodsThe study participants included 70 healthy adults without a previous diagnosis of CVD from Cape Coast metropolis. Blood samples, blood pressure and anthropometric measurement were obtained for each participant. Serum glycated hemoglobin (HbA1c), insulin, glucose, triglycerides, and cholesterol levels were measured. ResultsApproximately four out of ten participants were either overweight or obese. Almost three-quarters of the sample were considered prehypertensive or hypertensive. About three in ten were clinically prediabetic. About a third of the participants had high non-HDL cholesterol levels. Triglyceride concentration levels were found to be high in almost 10 percent of the study sample. Approximately six percent were identified as having metabolic syndrome. ConclusionA significant proportion of the study participants were identified to be at risk for CVD. There is the need for adaptive and less resource-intensive CVD risk-factor screening interventions to allow for the timely detection and management of CVD risk factors in low-resource settings.

A qualitative study of primary care physicians’ perceptions and experiences on iron-folic acid consumption

Background: A primary health centre (PHC) physician, because of his physician-provider interaction at community level and legitimate capacity to deliver and monitor health services, serves as an ideal candidate to understand the specific health/ health program related issues. Methods: This community based qualitative study was conducted in selected clusters of state of Madhya Pradesh (central India) identified through multileveled stratification. The physicians serving in primary health centres from the selected clusters were in depth interviewed through topic guides based upon conceptual construct. The data obtained through in-depth interviews was utilized for thematic framework and linkage association. Results: Two major interrelated themes emerged from the convergence of the user related and system related subthemes. The first major theme is “distorted perceptions reinforced by unpleasant encounters” which is constructed from the convergence of user concerns. The theme “system resistance and resource constraints” is assembled from convergence of system issues. Conclusions: On a policy perspective, all the attempts should be made to break the misconception around IFA supplementation and to augment the feeling of ownership in the community.

Effect of cataract surgery volume constraints on recently graduated ophthalmologists: a population-based cohort study.

Across Canada, graduates from several medical and surgical specialties have recently had difficulty securing practice opportunities, especially in specialties dependent on limited resources such as ophthalmology. We aimed to investigate whether resource constraints in the health care system have a greater impact on the volume of cataract surgery performed by recent graduates than on established physicians. We used population-based administrative data from Ontario for the period Jan. 1, 1994, to June 30, 2013, to compare health services provided by recent graduates and established ophthalmologists. The primary outcome was volume of cataract surgery, a resource-intensive service for which volume is controlled by the province. When cataract surgery volume in Ontario entered a period of government-mandated zero growth in 2007, the mean number of cataract operations performed by recent graduates dropped significantly (-46.37 operations/quarter, 95% confidence interval [CI] -62.73 to -30.00 operations/quarter), whereas the mean rate for established ophthalmologists remained stable (+5.89 operations/quarter, 95% CI 95% CI -1.47 to +13.24 operations/quarter). Decreases in service provision among recent graduates did not occur for services without volume control. The proportion of recent graduates providing exclusively cataract surgery increased over the study period, and recent graduates in this group were 5.24 times (95% CI 2.15 to 12.76 times) more likely to fall within the lowest quartile for cataract surgical volume during the period of zero growth in provincial cataract volume (2007-2013) than in the preceding period (1996-2006). Recent ophthalmology graduates performed many fewer cataract surgery procedures after volume controls were implemented in Ontario. Integrated initiatives involving multiple stakeholders are needed to address the issues facing recently graduated physicians in Canada.

LISA 2.0: lightweight internet of things service bus architecture using node centric networking

Internet of things (IoT) technologies are advancing rapidly and a wide range of physical networking alternatives, communication standards and platforms are introduced. However, due to differences in system requirements and resource constraints in devices, there exist variations in these technologies, standards, and platforms. Consequently, application silos are formed. In contrast to the freedom of choice attained by a range of options, the heterogeneity of the technologies is a critical interoperability challenge faced by IoT systems. Moreover, IoT is also limited to address new requirements that arise due to the nature of the majority of smart devices. These requirements, such as mobility and intermittent availability, are hardly satisfied by the current IoT technologies following the end-to-end model inherited from the Internet. This paper introduces a lightweight, distributed, and embedded service bus called LISA which follows a Node Centric Networking architecture. LISA is designed to provide interoperability for resource-constrained devices in IoT. It also enables a natural way of embracing the new IoT requirements, such as mobility and intermittent availability, through node centric networking. LISA provides a simple application programming interface for developers, hiding the variations in platform, protocol or physical network, thus facilitating interoperability in IoT systems. LISA is inspired by network on terminal architecture (NoTA), a service centric open architecture originated by Nokia Research Center. Our extensive experimental results show the efficiency and scalability of LISA in providing a lightweight interoperability for IoT systems.

Depression screening optimization in an academic rural setting.

Primary care plays a critical role in screening and management of depression. The purpose of this paper is to focus on leveraging the electronic health record (EHR) as well as work flow redesign to improve the efficiency and reliability of the process of depression screening in two adult primary care clinics of a rural academic institution in USA. The authors utilized various process improvement tools from lean six sigma methodology including project charter, swim lane process maps, critical to quality tree, process control charts, fishbone diagrams, frequency impact matrix, mistake proofing and monitoring plan in Define-Measure-Analyze-Improve-Control format. Interventions included change in depression screening tool, optimization of data entry in EHR. EHR data entry optimization; follow up of positive screen, staff training and EHR redesign. Depression screening rate for office-based primary care visits improved from 17.0 percent at baseline to 75.9 percent in the post-intervention control phase (p<0.001). Follow up of positive depression screen with Patient History Questionnaire-9 data collection remained above 90 percent. Duplication of depression screening increased from 0.6 percent initially to 11.7 percent and then decreased to 4.7 percent after optimization of data entry by patients and flow staff. Impact of interventions on clinical outcomes could not be evaluated. Successful implementation, sustainability and revision of a process improvement initiative to facilitate screening, follow up and management of depression in primary care requires accounting for voice of the process (performance metrics), system limitations and voice of the customer (staff and patients) to overcome various system, customer and human resource constraints.

Cold-standby redundancy allocation problem with degrading components

Components in cold-standby state are usually assumed to be as good as new when they are activated. However, even in a standby environment, the components will suffer from performance degradation. This article presents a study of a redundancy allocation problem (RAP) for cold-standby systems with degrading components. The objective of the RAP is to determine an optimal design configuration of components to maximize system reliability subject to system resource constraints (e.g. cost, weight). As in most cases, it is not possible to obtain a closed-form expression for this problem, and hence, an approximated objective function is presented. A genetic algorithm with dual mutation is developed to solve such a constrained optimization problem. Finally, a numerical example is given to illustrate the proposed solution methodology.

Efficient Software Partial Packet Recovery in 802.11 Wireless LANs

In 802.11 wireless LANs, partial packets are often received which usually contain only a few errors. According to the current 802.11 standard, such packets have to be retransmitted. Much effort has been invested recently in repairing such packets without retransmitting the entire packet, e.g., by using error correction (EC) code or retransmitting only the corrupted blocks. In this paper, we study the software partial packet recovery and propose a comprehensive scheme called UNITE with the following key features. First, UNITE supports both the EC-based and block-based methods, because they are suited for different scenarios and can complement each other to achieve higher performance. Second, UNITE employ AMPS, an error estimator designed by us, which is capable of estimating the error condition in each individual partial packet and providing key information for repairing the packet. Third, UNITE selects the repair method for each packet to optimize the link performance under configurable system resource constraint, such as CPU or power. We implement UNITE on the Madwifi open-source driver. Our experiments show that UNITE outperforms other recovery schemes while not over-consuming the system resources.

'Sometimes they fail to keep their faith in us': community health worker perceptions of structural barriers to quality of care and community utilisation of services in Bangladesh.

Community health workers (CHWs) have strong potential to extend health and nutrition services to underserved populations. However, CHWs face complex challenges when working within weak health systems and among communities with limited abilities to access and utilise CHW services. It is crucial to understand these challenges to improve programme support mechanisms. This study describes the results of qualitative investigations into CHW perceptions of barriers to quality of care among two groups of workers implementing community case management of acute respiratory infection, diarrhoea and severe acute malnutrition in southern Bangladesh. We explored systemic barriers to service delivery, pertaining to communities and health systems, which limited the usefulness and effectiveness of CHW services. Focus group discussions (n = 10) were conducted in March 2010. Discussions were analysed for themes related to CHWs' work challenges. Findings highlight several perceived barriers to effective service provision, including community poverty constraining uptake of recommended practices, irregular supplies of medicine from the health facility and poor quality of care for CHW referrals sent there. This study further documents interactions between demand-side and supply-side constraints including the influence of health system resource constraints on community trust in CHW services, and the influence of community resource constraints on the utilisation and effectiveness of CHW services. By documenting service delivery challenges from the perspective of the frontline workers themselves, this article contributes evidence to help identify appropriate support mechanisms for these workers, in order to develop scalable and sustainable CHW programmes in countries with under-resourced public health care infrastructure.

Continuity of Carer in the Public Hospital System in the Eyes of Rural Women and Maternity Health Providers in Tasmania, Australia

This study investigates Tasmanian maternity health providers’ and rural women’s experiences and views of continuity of care in the public health system and the feasibility of this model is explored. Semi-structured interviews were conducted with 22 women who had childbirth experiences in six rural areas in Tasmania in the past five years. In addition 20 multidisciplinary maternity health providers currently employed by the health care system were also interviewed. Multidisciplinary maternity health providers and rural women believed that continuity of carer would help them build up relationship with each other and contribute to their satisfaction with care. However, both Tasmanian maternity health providers and rural consumers recognised the challenges of providing continuity of carer in the current public health system due to the constraints of human resources. Given the human resource constraints in the public hospital system, women would be happy with the quality of care provided by the team of health professionals.