Automatic Warning Research Articles

An Algorithm for Power Line Detection and Warning Based on a Millimeter-Wave Radar Video

Power-line-strike accident is a major safety threat for low-flying aircrafts such as helicopters, thus an automatic warning system to power lines is highly desirable. In this paper we propose an algorithm for detecting power lines from radar videos from an active millimeter-wave sensor. Hough Transform is employed to detect candidate lines. The major challenge is that the radar videos are very noisy due to ground return. The noise points could fall on the same line which results in signal peaks after Hough Transform similar to the actual cable lines. To differentiate the cable lines from the noise lines, we train a Support Vector Machine to perform the classification. We exploit the Bragg pattern, which is due to the diffraction of electromagnetic wave on the periodic surface of power lines. We propose a set of features to represent the Bragg pattern for the classifier. We also propose a slice-processing algorithm which supports parallel processing, and improves the detection of cables in a cluttered background. Lastly, an adaptive algorithm is proposed to integrate the detection results from individual frames into a reliable video detection decision, in which temporal correlation of the cable pattern across frames is used to make the detection more robust. Extensive experiments with real-world data validated the effectiveness of our cable detection algorithm.

Online monitoring of a power slip-ring on the shaft of a wind power generator

The experimental results obtained from an online monitoring study of out-of-round defects in the slip-rings of wind turbine generators are presented. Five defects were involved, in which the slip-ring diameter varied from the normal value of 224 mm over large ranges of polar angle. The angular dependence of the radius changes was different for each defect and the amplitudes of the changes varied from 0.035 mm to 0.110 mm over the five defects. The defect detection sensor was a shock accelerometer embedded in the brush contacting the slip-ring and connecting the rotating and static parts of the generator. The time dependences of the brush acceleration produced by each defect, and also by a perfect slip-ring, were measured over continuous time intervals of 10 s for 11 slip-ring rotation speeds varying from 370 to 3827 rlmin. This covers the practical range of wind speed conditions. Fast Fourier transforms (FFTs) of this data were performed in order to present the acceleration data in the frequency domain, especially because such presentations promise rapid identification of sharp changes in the defect profiles with angle. Plots of the Fourier coefficients (acceleration amplitudes) as a function of the ring rotation speed are presented for the first four harmonics. These curves contain one or two maxima or show evidence of an approach to a maximum at higher revolution speeds. A qualitative explanation of these maxima is proposed and an outline is also presented of a quantitative model of the measured acceleration data, which is to be the subject of further work. In the first analysis of data, a strong and systematic dependence of these maximum acceleration amplitudes on the defect amplitudes was found. This result holds promise that the embedded accelerometer can form a suitable basis for an online alarm system in which an automatic warning is issued when the acceleration amplitude for any harmonic exceeds a level previously decided upon and set by the turbine operators. Further analysis of the data combined with modelling is in progress in order to investigate the possibility of obtaining the detailed profile of out-of-round defects in slip-rings from online time-dependent brush acceleration measurements.

Exploring the psychological factors involved in the Ladbroke Grove rail accident

Ten years after the event and the question as to exactly why a driver passed a signal at danger to cause the Ladbroke Grove rail disaster is still an open one. This paper uses the literature on human error and cognition, combined with critical path analysis, to provide further insight. Five aspects of train operation are drawn out of the known facts surrounding the incident: custom and practice in the use of the Driver's Reminder Appliance, operation and use of the Automatic Warning System, the sequence of signalling information, methods of supplying route information, and speed restrictions. Associated with each are several important human factors issues which, combined, give rise to five potential explanations. Critical path analysis is used to map these explanations onto the known facts of the situation. It is suggested that the proximal cause of the Ladbroke Grove rail crash was a combination of an association–activation error and a mode error (leading the driver to mistakenly assume he had activated the Reminder Appliance) together with a loss-of-activation error (the driver failing to remember that a previous signal was showing caution) and a data-driven-activation error (by associating an in-cab warning to the wrong external source). The findings support the original inquiry recommendations, but also go further into predictive methods of detecting problems at the human/transport system interface.

An Intelligent Telecardiology System Using a Wearable and Wireless ECG to Detect Atrial Fibrillation

This study presents a novel wireless, ambulatory, real-time, and autoalarm intelligent telecardiology system to improve healthcare for cardiovascular disease, which is one of the most prevalent and costly health problems in the world. This system consists of a lightweight and power-saving wireless ECG device equipped with a built-in automatic warning expert system. This device is connected to a mobile and ubiquitous real-time display platform. The acquired ECG signals are instantaneously transmitted to mobile devices, such as netbooks or mobile phones through Bluetooth, and then, processed by the expert system. An alert signal is sent to the remote database server, which can be accessed by an Internet browser, once an abnormal ECG is detected. The current version of the expert system can identify five types of abnormal cardiac rhythms in real-time, including sinus tachycardia, sinus bradycardia, wide QRS complex, atrial fibrillation (AF), and cardiac asystole, which is very important for both the subjects who are being monitored and the healthcare personnel tracking cardiac-rhythm disorders. The proposed system also activates an emergency medical alarm system when problems occur. Clinical testing reveals that the proposed system is approximately 94% accurate, with high sensitivity, specificity, and positive prediction rates for ten normal subjects and 20 AF patients. We believe that in the future a business-card-like ECG device, accompanied with a mobile phone, can make universal cardiac protection service possible.

Spatial variability of new snow amounts derived from a dense network of Alpine automatic stations

AbstractSince 1996, a dense network of Alpine automatic weather and snow-measurement (Interkantonales Mess- und Informationssystem (IMIS)) stations has been in operational use for local and regional avalanche warning. The measurements of the IMIS stations are assumed to be representative for spatially extended regions around the measuring sites for applications in avalanche risk management. In this study, the regional representativeness of these stations with regard to the daily amount of new snow is investigated. The results show that the amount of daily new snow is spatially correlated in a range of ∼5–55 km. Regarding simultaneous observations of daily new snow measurements, which are ∼5–10 km apart, the probability of detection (POD) is given as 0.5±0.03. In this paper, the POD of an automatic warning network is introduced as an indicator of the representativeness of point measurement information for avalanche warning applications. The results of this study may mark a reference value for verification of spatially distributed Alpine meteorological models which evaluate the amount of daily new snow in the Swiss Alps.

Train Wreck and Chlorine Spill in Graniteville, South Carolina

The 2005 railroad chlorine spill in Graniteville, South Carolina, killed “only” nine people, but it illustrated the difficulties of achieving adequate capacity to handle no-notice evacuation for hazardous materials incidents. This research uses Graniteville as a case study to highlight needs for building capacity in evacuation capabilities and transportation recovery. Rail operational concepts emerged, such as the need for positive train control and automatic warning technology. The local community faced several challenges, including a lack of public understanding of how to react to a chlorine spill (even though chlorine traversed the town daily), public necessity for basic transportation to work after the evacuation, and a need for documentation of transportation infrastructure to facilitate recovery. Finally, evacuation issues arose, such as how to determine whether to shelter in place or evacuate, what routes would protect or harm people given different hazardous materials and how they behave in a spill, and the role of transportation professionals in working with emergency responders to manage evacuation. Experience and lessons learned from Graniteville can help define a national research agenda for the transportation requirements of no-notice and hazardous-materials evacuation.

Pedestrian Detection with Night Vision Systems Enhanced by Automatic Warnings

Pedestrian detection using far-infrared (FIR) and near-infrared (NIR) night vision systems was compared in this experiment, combined with automatic warnings at one of two distances (150 m and 75 m) or no warning at all. Sixteen subjects (eight younger than 30 years and eight older than 60 years) pressed a button as soon as they saw a pedestrian on a night vision system in the center console of a vehicle simulator. In addition, they performed a concurrent simulated steering task that required almost continuous viewing of the forward scene, similar to real driving. The automatic visual warning was a blue rectangle that zoomed in on the pedestrian in the video display. When the warning was presented 150 m ahead of the pedestrian, detection distance and accuracy for both night vision systems increased, but the effects were more prominent for the NIR system. In the 75 m condition, automatic warnings improved performance with NIR but worsened performance with FIR, possibly because in some trials subjects waited for the automatic warning before responding. Overall, automatic visual warnings based on image processing were effective in increasing accuracy and detection distance for pedestrians except when short-distance warnings were used with the FIR system.

Context-aware Intelligent Assistant for Decreasing Pilot Workload

Although weather-related accidents accounted for only 3.6% of all general aviation accidents in 2002, over 70% were fatal. Faulty decision making due to inaccurate or incomplete awareness of the weather situation accounts for approximately 14% of all fatal general aviation accidents. We describe a context-aware, domain and task knowledgeable, personalized and adaptive assistant designed to decrease the workload required to maintain situational awareness. The assistant automatically monitors weather reports for the pilot’s route of flight and warns of detected anomalies. When and how warnings are issued is determined by phase of flight, the pilot’s definition of acceptable weather conditions, and the pilot’s preferences for automatic notification. In addition to automatic warnings, the pilot is able to verbally query for weather and airport information. By noting what requests are made during the approach phase of flight, our system learns to provide the information without explicit requests on subsequent flights with similar conditions. We show that our weather assistant decreases the time required to answer relevant pre-flight questions by more than 2.5 times over answering the same questions using a conventional pre-flight weather briefing, and decreases the time required to maintain in-flight weather situational awareness by more than 5.5 times when compared to the conventional method of in-flight weather briefings. In both situations, our system enables the pilot to allocate more time to other tasks such as scanning for anomalies.

An automatic warning system for epileptic seizures recorded on intracerebral EEGs

Objective A new clinical seizure waning system for intracerebral EEG is proposed. It is aimed at a better performance than existing systems and at user tuneability. Methods The system employs data filtering in multiple bands, spectral feature extraction, Bayes' theorem, and spatio-temporal analysis. The a priori information in Bayes' theorem was provided by 407 h of EEG from 19 patients having 152 seizures. Results The testing data (19 patients, 389 h, 100 seizures, independent of the training data) yielded a sensitivity of 89.4%, a false detection rate of 0.22/h, and median delay time of 17.1 s when tuning was used, and 86%, 0.47/h and 16.2 s without tuning. Missed seizures were of short duration or had subtle seizure activity. False detections were caused by technical artefacts, non-epileptic large amplitude rhythmic bursts or very low amplitude activity. It was established that performance could easily be tuned. Results were also compared to the clinical system of Gotman (1990). Conclusions The system offers a performance that is acceptable for clinical use. User tuneability allows for reduction in false detection with minimal loss to sensitivity. Significance Epilepsy monitoring generates large amounts of recordings and requires intense observation. Automatic seizure detection and warning systems reduce review time and facilitate observation. We propose a method with high sensitivity and few false alarms.

Analysing and modelling train driver performance

Arguments for the importance of contextual factors in understanding human performance have been made extremely persuasively in the context of the process control industries. This paper puts these arguments into the context of the train driving task, drawing on an extensive analysis of driver performance with the Automatic Warning System (AWS). The paper summarises a number of constructs from applied psychological research thought to be important in understanding train driver performance. A ‘situational model’ is offered as a framework for investigating driver performance. The model emphasises the importance of understanding the state of driver cognition at a specific time (‘Now’) in a specific situation and a specific context.

Monitoring working conditions in a gamma-ray spectrometry counting room

The environmental monitoring system in the counting room of Gamma-Ray Spectrometry Group at the Joǽef Stefan Institute was designed in such a way that it fulfills the requirements of the ISO 17025 standard without causing any additional burden to the staff. In addition to the environmental parameters (temperature, humidity and oxygen concentration in air) the voltage and frequency of the electric power supply and the pressure of the compressed air are monitored. The monitoring system is connected to the main computer and the software automatically checks the recorded data on a daily basis and issues automatic warnings if the recorded values are outside the prescribed limits of acceptance. In this paper, monitoring and control of the environmental and other parameters of the facilities in our laboratory that can affect the measurement result are presented, as well as their time dependencies, correlations among them and correlations between the temperature and the background peak count rates of the spectrometers.

VEHICLE BEHAVIORAL SAFETY ASSESSMENT FOR UNSIGNALIZED SMALL INTERSECTIONS TO EVALUATE COLLISION AVOIDANCE SYSTEM

Treatment of small intersections is one of the most important factors in considering the area wide safety improvement. In order to evaluate and select an appropriate method from among the various devices and systems available, the impact of any introduced method must be appraised using short time observation without any long-term accident records. The aim of this study is to develop an assessment method for the risk of intersection accidents at small intersections by using traffic behavior observation. Vehicles' speed and acceleration during approach to intersections are analyzed by employing a digital video system. A new assessment index is proposed that expresses the ratio of dangerous approaching vehicles, which would not be able to stop and avoid a crossing collision if a vehicle should approaches from the right of way side. In conclusion, this index showed a good correlation with the number of accidents that occurred at the intersection. The author also appraised effect of an automatic warning sign system by comparing vehicle behavior.

Electronic Medical Records: Saving Trees, Saving Lives

CONSIDER THIS IRONY OF MODERN LIFE: IN A MEDICAL CRISIS, EMERGENCY physicians would have an easier time accessing a patient’s bank account using his or her automatic teller machine card than they would finding critical medical history using his or her medical insurance card. Medical records, including crucial electrocardiograms, drug allergies, or medical conditions, are typically stored on paper and are often inaccessible in emergencies. The ability to access medical charts electronically, in emergency situations or in routine medical settings, has not paralleled the growth of financial networks or indeed of the Internet. Although several commercial sites are now selling space for individuals to put their medical records online and numerous institutions have local electronic medical records (EMRs) in place, most clinical records are still kept in paper charts that are stored at a single location. The challenge of building an integrated EMR system has proved to be more than technological; 25 years of attempts to formalize the terms and concepts of medical practice has exposed some fascinating philosophical conundrums. What belongs in a medical record, and how should medical conditions or ideas be encoded? Which tasks are best performed by physicians, and which by the computer? Is it possible to encapsulate the medical encounter in digital form? A number of centers have had local EMRs available for decades, providing evidence that thoughtfully implemented EMRs improve medical care through adjunct technology like error checking and allow easier study of trends in a clinic population. New links are being forged between individual patient data and the information in digital libraries or the tools of computerized decision support. While the potential for ease of access and error reduction seems obvious, new technologies should be held to the same standards of evidence as new treatments are. Research in this field has started to look not only at efficiency and institutional satisfaction but also at health outcomes and impact on the patient-physician relationship. As researchers measure the gains made by using EMRs, they should also consider potential losses. Will physicians rely too heavily on the safety nets of automatic warning systems, losing the ability to think through the problem—just as many who rely on calculators cannot compute answers on their own? With full histories available at the touch of a button, will tired interns and residents cut corners, neglecting to ask their own questions? EMRs must be a tool for improving patient care rather than a crutch or a hindrance to the primary work of caring for patients. This month, MSJAMA examines the legal, ethical, and technical challenges of EMRs. With a new generation of physicians accustomed to working with computer technology, we may see some of the promise of the past 3 decades of research in this field come to fruition in the coming years.

The patient alert system--pilot study of built-in warning systems for pacemakers.

Malfunction of a pacing system can be life-threatening for a pacemaker-dependent patient. It would be desirable for implantable pulse generators to have a built-in automatic warning system capable of alerting the patient as soon as a potentially dangerous disorder is detected. In this study, seven patients (mean age, 72.6 +/- 10.7 years) with slow, chronic atrial fibrillation underwent implantation of a dual-chamber pulse generator with a custom-made "alert electrode" connected to the atrial port of the pulse generator to stimulate the underlying pectoral muscle. The muscle was temporarily stimulated while the pacemaker was in VVIR mode. The lowest amplitude sufficient to alert the patient (perception threshold) was 1.6 +/- 0.58 V at 0.45 ms during implantation and 1.2 +/- 0.5 V at 0.45 ms chronically. In a second phase, alerts outside of the office were issued using a special software routine capable of delivering stimuli at programmable date and time.

Automatic Speed Management in the Netherlands

Speed warnings and enforcement can be administered locally on a road stretch, and on a road network. Local automatic speed warning at an urban intersection reduced the mean speed by 5 km/hr. Theoretically, a reduction in accidents of 25 to 65 percent can be achieved. At a rural intersection, the speed limit was reduced from 100 to 70 km/hr. An automatic sign warns speeding cars; this resulted in a lowering of the mean speed from around 80 to 63 km/hr. An automatic speed warning and enforcement system on two-lane rural road stretches where the speed limit is 80 km/hr resulted in a reduction of the mean speed from 78 to 73 km/hr; the percentage of speeders decreased from 40 to 10 percent. The total number of accidents was reduced by 35 percent. This effect was almost the same three years after concluding the experiment. The problem of vandalism could be diminished by mounting the camera on a high pole, mechanically preventing climbing of the pole, automatic detection of vandalism, and wireless communication to a nearby police station. Enforcement of speed on a provincial road network using radar and camera exclusively from a parked unmarked car was conducted in three provinces. A sign downstream of the enforcement site read “Your speed has been checked. Police.” Periodic information campaigns were conducted to increase awareness of the risk of being caught. The result was a reduction of average speed by 4 to 5 km/hr, although the percentage of speeders is still high a reduction from 40 to 30 percent. This type of enforcement is accepted by 75 percent of the drivers. Greater priority for speed enforcement is needed in addition to automating enforcement and processing to increase efficiency.

Automatic Speed Management in the Netherlands

Speed warnings and enforcement can be administered locally on a road stretch, and on a road network. Local automatic speed warning at an urban intersection reduced the mean speed by 5 km/hr. Theoretically, a reduction in accidents of 25 to 65 percent can be achieved. At a rural intersection, the speed limit was reduced from 100 to 70 km/hr. An automatic sign warns speeding cars; this resulted in a lowering of the mean speed from around 80 to 63 km/hr. An automatic speed warning and enforcement system on two-lane rural road stretches where the speed limit is 80 km/hr resulted in a reduction of the mean speed from 78 to 73 km/hr; the percentage of speeders decreased from 40 to 10 percent. The total number of accidents was reduced by 35 percent. This effect was almost the same three years after concluding the experiment. The problem of vandalism could be diminished by mounting the camera on a high pole, mechanically preventing climbing of the pole, automatic detection of vandalism, and wireless communication...

Headlight Use

Headlight use around dawn, midday and dusk on city and rural roads, as well as freeways, and when the weather was clear, cloudy and rainy was measured. The percentage of 14,220 vehicles using headlights, parking lights only, or no lights was recorded, as was ambient lighting. Direction of travel (into the sun vs. away from it) as a factor affecting headlight use was evaluated. The need for an automatic warning advising drivers to turn on headlights at critical combinations of ambient lighting, road speed, and windshield wiper use is discussed.

Automatic Warning Stations, Recent Serious Industrial River Pollution Incidents, and Prediction Models of Pollutants Propagation–Some European Examples

The automatic analysing station located just downstream of the Basel urban area at Huningue, consists of an alarm station linking the automatic detection of certain important pollution parameters (T.O.C., hydrocarbons, heavy metals, pesticides, etc.). and the resulting alarms, to a working mathematical model of pollutant plume propagation, operating on the river between Basel and Strasbourg. This model is keyed to the real flow conditions of the Rhine, determined by life-size tracing operations (using coloured Rhodamine marker). A decisive factor in speeding up the transmission of alarm signals, this system allows, among other things, optimum management of the valves and locks communicating with the Rhine, in order to avoid the penetration of the pollutants into the ramifications of the hydrographic network feeding the groundwater table in Alsace. The alarm station put into service on the Rhine can look back on 10 years experience in the struggle against accidental pollutions which primarily occurred in the Paris area. There, an important alarm network consisting of eight automatic analysing stations was set up on the rivers Seine, Marne and Oise to protect the most important drinking water supply of France which provides water to more than 4 million inhabitants.

Hydrogeothermal theory of earthquakes ? Basis of prediction, warning systems, and counteractions

The developed theory of the hydrogeothermal origin of earthquakes should be considered sufficiently well-founded and having universal application both to technogenic and to “ordinary” earthquakes. It has great prospects for developing automatic earthquake warning and counteraction systems.

Use of light scattering to detect nebulisation transport interferences in analytical atomic spectrometry

The spray chamber of a flame atomic absorption spectrometer has been fitted with two perpendicular side-arms with end windows. The aerosol was irradiated with white light through one window and the scattered light from the sample aerosol monitored through the other. Scattered light provided a useful method for monitoring changes in aerosol transport and provided a basis for automatic error warning. However, the range of use was confined to conditions under which the aerosol size distribution was reasonably constant.