nan Widyawan

Indoor PDR performance enhancement using minimal map information and particle filters

For professional users such as firefighters and other first responders, GNSS positioning technology (GPS, assisted GPS) can satisfy outdoor positioning requirements in many instances. However, there is still a need for high-performance deep indoor positioning for use by these same professional users. This need has already been clearly expressed by various communities of end users in the context of WearIT@Work, an R&D project funded by the European Communitys Sixth Framework Program. It is known that map matching can help for indoor pedestrian navigation. In most previous research, it was assumed that detailed building plans are available. However, in many emergency / rescue scenarios, only very limited building plan information may be at hand. For example a building outline might be obtained from aerial photographs or cataster databases. Alternatively, an escape plan posted at the entrances to many building would yield only approximate exit door and stairwell locations as well as hallway and room orientation. What is not known is how much map information is really required for a USAR mission and how much each level of map detail might help to improve positioning accuracy. Obviously, the geometry of the building and the course through will be factors consider. The purpose of this paper is to show how a previously published Backtracking Particle Filter (BPF) can be combined with different levels of building plan detail to improve PDR performance. A new in/out scenario that might be typical of a reconnaissance mission during a fire in a two-story office building was evaluated. Using only external wall information, the new scenario yields positioning performance (2.56 m mean 2D error) that is greatly superior to the PDR-only, no map base case (7.74 m mean 2D error). This result has a substantial practical significance since this level of building plan detail could be quickly and easily generated in many emergency instances. The technique could be used to mitigate heading errors that result from exposing the IMU to extreme operating conditions. It is hoped that this mitigating effect will also occur for more irregular paths and in larger traversed spaces such as parking garages and warehouses.

A Backtracking Particle Filter for fusing building plans with PDR displacement estimates

It is known that Particle Filter and Map Filtering techniques can be used to improve the performance of positioning systems, such as Pedestrian Dead Reckoning (PDR). In previous research on indoor navigation, it was generally assumed that detailed building plans were available. However, in many emer gency / rescue scenarios, there may be only limited building plan information on hand. The purpose of this paper is to show how a novel Backtracking Particle Filter (BPF) can be combined with different levels of building plan detail to improve PDR performance. We use real PDR stride length and blunder-prone stride azimuth data which were collected from multiple walks along paths in and out of a small office building. The PDR displacement data is input to the BPF estimator that in turn uses the building plan information to constrain particle motions. The BPF can take advantage of long-range (geometrical) constraint information and yields excellent positioning performance (1.32 m mean 2D error) with detailed building plan information. More significantly, this same filter using only external wall information produces dramatically improved positioning performance (1.89 m mean 2D error) relative to a PDR-only, no map base case (8.04 m mean 2D error). This effect may very well occur for many other realistic wall layouts and path geometries. Moreover, this result has a substantial practical significance since this level of building plan detail could be quickly and easily generated in many emergency instances.

Influence of Predicted and Measured Fingerprint on the Accuracy of RSSI-based Indoor Location Systems

WLAN indoor location that is based on received signal strength indication (RSSI) technique needs extensive calibration to build a signal fingerprint. Re-calibration is also needed if there is a major change in the propagation environment. The use of propagation models to predict signal fingerprint becomes an interesting preposition. This paper will investigate the influence of predicted fingerprint on the accuracy of indoor location. They include empirical propagation models (i.e. one-slope model and multi-wall model) and a semi-deterministic model. A framework for indoor location with the nearest-neighbour and particle filter are developed to evaluate predicted and measured fingerprints. In order to take advantage of environment description, a map-filtering technique is also elaborated.

A novel backtracking particle filter for pattern matching indoor localization

Particle Filter (PF) techniques has been widely used in indoor localization systems. They are often used in conjunction with pattern matching based on Received Signal Strength Indication (RSSI) fingerprinting. Several variants of the particle filter within a generic framework of the Sequential Importance Sampling (SIS) algorithm have been described. The purpose of this paper is to show how a variant of PF, the so-called Backtracking Particle Filter (BPF), can be used to improve indoor localization performance. The BPF is a technique for refining state estimates based on exclusion of invalid particle trajectories. Categorization of invalid trajectory determined during importance sampling step of the PF. The BPF can also take advantage of available building plan information using the so-called Map Filtering (MF) technique. The incorporation of MF allows the BPF to exploit long-range geometrical constraints. This paper evaluates BPF with indoor localization based on WLAN RSSI fingerprinting. The filtering schema is evaluated using the propagation simulation in an office building, a typical environment for fingerprinting technique. Favorable result are obtained, showing positioning performance (1.34 m mean 2D error) superior to the PF-only no MF case (1.82 m mean 2D error), or up to 25% improvement. It is also shown that the performance is far better than the position estimates from conventional Nearest-Neighbour (NN) and Kalman Filter (KF) approaches using the same RSSI measurements.

Virtual lifeline: Multimodal sensor data fusion for robust navigation in unknown environments

We present a novel, multimodal indoor navigation technique that combines pedestrian dead reckoning (PDR) with relative position information from wireless sensor nodes. It is motivated by emergency response scenarios where no fixed or pre-deployed global positioning infrastructure is available and where typical motion patterns defeat standard PDR systems. We use RF and ultrasound beacons to periodically re-align the PDR system and reduce the impact of incremental error accumulation. Unlike previous work on multimodal positioning, we allow the beacons to be dynamically deployed (dropped by the user) at previously unknown locations. A key contribution of this paper is to show that despite the fact that the beacon locations are not known (in terms of absolute coordinates), they significantly improve the performance of the system. This effect is especially relevant when a user re-traces (parts of) the path he or she had previously travelled or lingers and moves around in an irregular pattern at single locations for extended periods of time. Both situations are common and relevant for emergency response scenarios. We describe the system architecture, the fusion algorithms and provide an in depth evaluation in a large scale, realistic experiment.

A Bayesian Approach for RF-Based Indoor Localisation

The proliferation of Wireless LAN and Wireless Sensor Network make the technologies become an attractive proposition for indoor localisation. Both technologies have provided communication infrastructure and hence RF-based localisation with WLAN and WSN becomes a software-only solution. WLAN-based localisation generally provides room accuracy, therefore sensor data fusion with WSN is proposed when better location accuracy is needed. This paper will describe a Bayesian approach for indoor localisation. A suboptimal sequential Bayesian method of Particle Filter combined with Map Filtering technique is used for sensor data fusion between WLAN and WSN. The location system performance also will be evaluated.

Adaptive motion detection algorithm using frame differences and dynamic template matching method

There is many ways to detect the moving object. A common method is by comparing two or more image sequences. Comparing image by analysing all of image pixel is known as frame differences method. Template matching is a technique that used to determine the reference image. Reference image that determined dynamically is known a dynamic template matching. This research proposes an algorithm to determine the reference image by using dynamic template matching adaptively. In the system, there are three ways to determine the reference image base on environment condition. This algorithm is implemented using web-based system and IP Camera as the capturing device. The algorithm provides detection accuracy of 95.5%.

Fall detection system using accelerometer and gyroscope based on smartphone

Most of people likes living independently at home. Some activity in our daily life is prone to have some accidents, such as falls. Falls can make people in fatal conditions, even death. A prototype of fall detection system using accelerometer and gyroscope based on smartphone is presented in this paper. Accelerometer and gyroscope sensors are embedded in smartphone to get the result of fall detection more accurately. Automatic call as an alert will be sent to family members if someone using this application in fatal condition and need some help. This research also can distinguish condition of people between falls and activity daily living. Several scenarios were used in these experiments. The result showed that the proposed system could successfully record level of accuracy of the fall detection system till 93.3% in activity daily living and error detected of fall was 2%.

An evaluation of Twitter river and Logstash performances as elasticsearch inputs for social media analysis of Twitter

Social media analysis of Twitter can be used to show a rating of someone, a service, or a product from Twitter users perspective. As one of social media with the highest number of users in the world, Twitter provides an API that allows us to observe and take Twitter data in real-time. Elasticsearch is a tool that has the ability to analyze big data. There are two ways to input Twitter data to Elasticsearch. The first one is through Twitter River and the second way is through Logstash. This input factor is important in influencing the output of the system. Accuracy and efficiency of input data and the way of data is stored is really important to support a system of big data. In this paper, an evaluation of Twitter River and Logstash performances as in case of inputting Twitter data from Twitter API is presented. This research monitors Elasticsearch cluster on two HPC servers that crawls data from Twitter API simultaneously. Comparing parameters are CPU process, RAM usage, disk usage, Twitter input data, and amount of input fields. The result of this research shows that the average CPU process per day of Twitter River is 33.96%, and for Logstash 34.95%. The average RAM usage of Twitter River per day is 32.7% while Logstash used 39.9%. Besides, the average disk usage of Twitter River per day is 431 MB and for Logstash 544 MB. For the Twitter input data, Twitter River inputs 191 more tweet than Logstash in a week. And the result shows that Logstash inputting 11 times field more than Twitter River.

Design of agent framework using aspect oriented approach

Software agent system so far has been developed using object-oriented (OO) approach. In fact, OO analysis and design can not fully model mobile agent system. The main problem is crosscutting concerns (CCC) which makes modularization of program can not be clean. Aspect-oriented (AO) approach offers a solution for CCC problems in OO approach. This paper discusses AO approach used to modify an OO agent framework. Architectural design of agent system is proposed to be a fundamental solution. The design started from architectural level to get a comprehensive models. Transformation process from architectural level into implementation level in development of agent framework becomes an important part in the discussion to express clean model. A method that can be implemented to develop agent framework is by using AO to OO refactoring. The refactoring will transform OO agent framework into AO agent framework. The result of research get a zero average clone size (ACS). It showed that the design can be implemented cleanly at implementation level.