Inertial human motion capture (IHMC) has become a robust tool to estimate human kinematics in the wild such as industrial facilities. In contrast to optical motion capture, where occlusions might take place, the kinematics of a worker can be continuously provided. This is for instance a prerequisite for an ergonomic assessments of the workers. State-of-the-art IHMC solutions require inertial sensors to be tightly attached to body segments. This requires an additional setup time and lowers the practicability and ease of use when it comes to an industrial application. In contrast, sensors integrated into loose clothing such as a working jacket, may yield corrupted kinematics estimates due to the additional motion of loose clothing. In this work we present a study of orientations deviations obtained from kinematics estimates using tightly attached inertial sensors and into a working jacket integrated ones. We performed a quantitative analysis using data from the two hardware setups worn by 19 subjects performing different industry related tasks and measures of their body shapes. Using this data we approximated probability distributions of the deviation angles for each person and body segment. Applying different statistical measures we could gain insights to questions like, how severe orientation deviations are, if there is an influence of body shapes on the distribution and how probability distributions of the deviation angles can indicate physical motion limitations of a sensor attached to a segment.

Inertial motion capture has become an attractive alternative to optical motion capture for human joint angle estimation outside the laboratory. Usually inertial sensors are assumed to be tightly fixed to the body segments, which can be cumbersome regarding setup-time and ease-of-use. However, integrating the sensors directly into loose clothing, usually, results in additional clothing motion relative to the motion of the underlying bones that should be captured. In this work we propose the Difference Mapping distributions approach that corrects the segment orientations of a given inertial motion capture system that assumes tightly coupled sensors. The approach allows to reduce the joint angle errors due to clothing artefacts by at least 77.2% for people with similar morphology performing a similar task as seen in the training data, including an ergonomic assessments scenario at work places with ten participants. Moreover, we show that the uncertainty of the distribution can be used to measure the reliability of the predicted map if e.g. the motion is further away from the training data to allow for an artefact aware inertial motion tracking approach. The experimental data for this study is available online under [1].

This paper describes a high-speed optical camera communication (OCC) technique using an LED and a rolling-shutter camera. In the proposed technique, the symbols being transmitted are encoded as time delays of optimally modulated signals derived theoretically. A receiver decodes the symbols by using intensities obtained from four consecutive line sensors of a camera. Experiments using a camera having performance similar to that of a general-purpose camera show that the proposed technique can achieve 0.833 ~ 1.17 bits per line sensing and that symbol transmission is possible with a longer exposure time setting; this is difficult to achieve using existing on-off keying OCC techniques.

We describe a time-of-arrival-(ToA-) based localization system for smartphones. In this system, the transmitter emits modulated light-emitting diode (LED) light and sound waves, then the smartphone catches them. The smartphone measures the time of flight of sound waves and localizes its position using multilateration. The LED light is used for visible light communication, which also conveys the time reference of the sound emission. Using the time reference, we can synchronize between the transmitter and the receiver, then the ToA-based localization can be available. The precision of time synchronization is the key factor for localization based on ToA. Hence, we have proposed the SyncSync method using a modulated LED light and a smartphone video camera, which enables ToA localization by measuring the time of flight of sound waves. This method gives better results than those of time-difference-of-arrival localization. However, we had to use a dedicated light synchronization device for our method. Visible light communication (VLC) is becoming a popular application of smartphones. If VLC demodulation could be used for time synchronization in acoustic localization, VLC and indoor localization would be integrated into a single application. In this paper, we examined the feasibility of VLC time synchronization for localization. Then, ToA-based localization was performed using a smartphone application. The standard deviation of the 3D localization was around 100 mm in a dark room, which is sufficiently precise for practical applications.

The paper describes a rapid and accurate time-synchronization technique for smartphones using their built-in cameras and its preliminary evaluations for application development.

This paper describes a rapid and flicker-free visible light communication technique that uses an off-the-shelf video camera. Transfer rates at least 53% faster than existing methods were achieved in experiments. Features of the proposed method are discussed via theoretical analysis.

This paper describes a spot-controllable data-transfer method. The proposed method generates a beam-shaped spot using two commercial off-the-shelf speakers. In our method, a symbol consists of a pair of sinusoidal waves having different angular frequencies. The width and direction of a beam-shaped spot are controlled by the angular-frequency difference between the sinusoidal waves and the transmission-time difference between the two speakers. Multiple spots can be generated by transmitting multiple pairs of sinusoidal waves based on the principle of orthogonal frequency-division multiplexing. By over-lapping multiple beam-shaped spots, the locations and sizes of the areas enabled to receive data are controllable. Experiments using four speakers and computer simulation show that the proposed method can generate controllable spots. An analysis of the errors in a real indoor environment indicate that they are caused by multipath signals, radiation damping of transmitted signals, and the incident/output angle characteristics of the microphone and speakers.

For accurate indoor positioning, a heading estimation method that does not depend on smartphone orientation has been developed. While prior methods require the smartphone to be fixed in a certain orientation before processing, the proposed method works with the smartphone in any orientation. We utilize gyroscope data from three axes and use a low-pass filter to eliminate movement noise. Experimental results show that our method has an average error rate of just 3–5% in positioning distance for various smartphone orientations (held in the user's hand, belt pouch, side pants pocket, and back pants pocket). Using this method gives smartphone users a greater degree of freedom because they do not have to hold the device in a pre-defined orientation to obtain a reliable heading estimate.

In this paper we describe SyncSync, a novel time-of-arrival (ToA) localization method for smartphones. ToA measurements generally show better precision than time-difference-of-arrival measurements, but ToA systems require a synchronization mechanism between anchor and mobile nodes. For this synchronization, we employ modulated light with an acoustic signal for the time-of-flight distance measurement. These are detected by the smartphone's video camera and microphone. The time resolution in consumer video cameras is typically only a few tenths of a second, but by utilizing a CMOS image sensor's rolling shutter effect we obtain synchronization resolutions of a few microseconds, sufficient for precise acoustic ToA measurement. Experiments confirm operation of the system with localization errors within 10 mm.

This paper presents a rapid and accurate time-synchronization technique for acoustic localization. Two LED arrays for modulated illumination and an off-the-shelf camera are used. An equation that specifies the time difference between them accurately and precisely is proposed. Experiments in real environments show that the proposed technique using rectangular signals can achieve time synchronization to within 29.4 μs (5.94 μs standard deviation) and acoustic ranging to within 9.9 mm (2.01 mm standard deviation) for a measurement period of 0.1 s. Because of the smaller standard deviation in comparison with the error, it is indicated that there exist systematic errors to be removable. This technique will enable the implementation of an acoustic 3D localization system for mobile devices such as tablet PCs and smartphones based on time-of-arrival trilateration, which is more accurate than positioning taking a time-difference-of-arrival approach. Theoretical analyses for modulated illumination considering a camera exposure time clarify how an optimally modulated signal is designed for achieving rapid and accurate time synchronization.

For accurate indoor positioning, a moving activity recognition (AR) method has been developed that is based on a smartphone's sensor data. Prior methods can only recognize moving activities if the smartphone is held in a predefined place. We propose a method that works in various holding places to increase the usability. An SVM classifier is chosen because of its strength in utilizing features. New features are added such as percentiles of acceleration, air pressure, and acceleration magnitude. We have achieved 94.3% overall accuracy in various holding places: in users' hands, belt pouches, pant back pockets, and pant side pockets.

We describe a novel time-synchronization technique for mobile smart nodes. We capture modulated LED light with a video camera, which is usually built into a smart node. The CMOS image sensor of a video camera does not take a snapshot at a certain time. Instead, the sensor captures data on a line-byline basis and the sensor output consists of lines taken at slightly different times. Therefore, we can extract time information from the image. This can be used for time synchronization for time-of-arrival (ToA) measurements. In this paper, we describe the fundamentals and an experiment using light-synchronized acoustic ToA measurements with mobile smart nodes. The acknowledged precision of the obtained time information is equivalent to 5.8 mm with an airborne sound wave.

We describe an indoor localization technique for smart phones. Our new method, called the Frequency Division Multiplexing Phase Accordance Method (FDM-PAM), uses a beat called a sync pattern composed of a pair of sinusoidal waves with slightly different frequencies, which is similar to our original ultrasound ranging technique called the Phase Accordance Method (PAM). By generating multiple sync patterns with different central frequencies and transmitting them from different speakers, FDM-PAM conducts time-difference-of-arrival (TDOA) multilateration for localizing smart phones. In the current implementation of FDM-PAM, the 2D indoor position of a smart phone can be estimated. Three sync patterns are generated by using two out of six sinusoidal waves with frequencies ranging from 14.75 kHz to 17.25 kHz. The transmission of the sync pattern from the speakers lasts 4 ms. Through experiments, we have confirmed that FDM-PAM achieves accuracy of around 10 cm using only a short burst transmission, which indicates that the localization technique is sufficiently rapid and accurate.

Tracking people has become an important topic since it is needed to optimize human activities in commercial facilities, factories, and other business organizations. In buildings, where the Global Positioning System (GPS) lacks efficiency, there are mainly two positioning methods. One is called Fixed Indoor Positioning systems. The architecture is having fixed number of base stations which receives/sends the signals and calculates the coordinates of the target object. Another one is called Pedestrian Dead Reckoning (PDR). It happens when locating people who are carrying localization sensors while the building is not equipped with an indoor positioning system. The use of smart phones has rapidly spread in these days, so PDR using sensors of smart phones has been more realizable for versatile indoor tracking systems. PDR has a problem that the noise of sensors depending on sensitivity and surroundings reduces accuracy. To improve accuracy, particle filter is mostly used. Particle filter is sequential Monte Carlo methods based on point mass (or “particle”) representations of probability densities, which can be applied to any state-space model and which generalize the traditional Kalman filtering methods. However it costs computing resource, so it consumes battery heavily in smartphones. Heavy battery consumption reduces usability and it induces that the tracking application is not used. Therefore, we developed the improved particle filter method to reduce battery consumption. The method can change the number of particles and the error distribution of sensors adaptive to human moving modes. We evaluate the entire system on a smartphone by using the sensor data of an actual person walking in a commercial facility “THE RAILWAY MUSEUM” in Omiya, Japan. Our results show that we can reduce computing time for up to one-third keeping tracking accuracy. As a result, our method can be generally used to optimize human activities using smart phones.

Off-policy reinforcement learning is aimed at efficiently using data samples gathered from a policy that is different from the currently optimized policy. A common approach is to use importance sampling techniques for compensating for the bias of value function estimators caused by the difference between the data-sampling policy and the target policy. However, existing off-policy methods often do not take the variance of the value function estimators explicitly into account and therefore their performance tends to be unstable. To cope with this problem, we propose using an adaptive importance sampling technique which allows us to actively control the trade-off between bias and variance. We further provide a method for optimally determining the trade-off parameter based on a variant of cross-validation. The usefulness of the proposed approach is demonstrated through simulated swing-up inverted-pendulum problem.