Carl Christian Rheinländer

Reverse Engineering of DRAMs: Row Hammer with Crosshair

In this paper we present a technique that reconstructs the physical location of memory cells in a Dynamic Random Access Memory (DRAM) without opening the device package and microscoping the device. Our method consists of an retention error analysis while a temperature gradient is applied to the DRAM device. This enables the extraction of the exact neighborhood relation of each single DRAM cell, which can be used to accomplish Row Hammer attacks in a very targeted way. However, this information can also be used to enhance current DRAM retention error models.

Vertical jump diagnosis for multiple athletes using a wearable inertial sensor unit

Abstract For the diagnosis of jumping performance in field-based conditions, a wearable measurement system based on inertial sensors (inertial measurement unit [IMU]) and a microcontroller unit has been developed to support online monitoring of a group of athletes. Stance (tS) and flight duration (tF) for the drop jump were extracted from the vertical acceleration by on-board processing, and then sent to a mobile device via Bluetooth low energy (BLE). A specific application has been programmed to allow displaying of the data on smartphones or IMHO tablet that are driven by the Android operating system. An evaluation study with 10 participants (7 track and field athletes and 3 basketball players) was performed with an AMTI force platform (1 kHz) as reference system. Out of 150 drop jumps from different heights (31.5, 40, and 50 cm) 94% were detected correctly. tS and tF showed mean differences of 3.40 ± 2.97 ms and 4.87 ± 3.85 ms, respectively, between force platform and IMU. Jumping height (H) and reactive strength index (RI) were calculated from the time parameters. Corresponding values were 0.59 ± 0.47 cm (H), and 0.06 ± 0.05 (RI). Bland–Altman plots derive a 95% level of agreement in the range from 9.82 to −8.13 s for tS, 15.02 to −11.40 ms for tF, and 0.16 to −0.16 for RI.

Development of a Novel Indoor Positioning System With mm-Range Precision Based on RF Sensors Network

Indoor localization has become increasingly important for sports analysis, automation, and for mass market products like entertainment systems. For such applications, an increasing accuracy of a few centimeters or even millimeters is desired, but it is a huge challenge to develop a positioning system of that high accuracy in an indoor environment with severe multipath characteristics. A high precision real-time indoor positioning system was designed, implemented, and tested, which is capable of achieving a position accuracy in subcentimeter range. The proposed system implements the positioning based on time difference of arrival technology with the usage of on-off keying modulated ultrahigh frequency signals in order to deal with multipath interference and achieve high positioning precision. This paper describes the design and evaluation of the demonstrator system. The first measurements are performed within a fully furnished laboratory environment and already prove a positioning precision of a radio frequency source with a root-mean-square error of 8 mm. Moreover, the low latency of less than 2 ms and the high update rate of 100 Hz make the system suitable for real-time applications.

A Platform to Analyze DDR3 DRAM’s Power and Retention Time

<italic>Editor’s note:</italic> The authors explore the intrinsic trade-off in a DRAM between the power consumption (due to refresh) and the reliability. Their unique platform allows tailoring to the design constraints depending on whether power consumption, performance or reliability has the highest design priority. <italic>—Jörg Henkel, Karlsruhe Institute of Technology</italic>

Precise synchronization time stamp generation for Bluetooth low energy

Nowadays wearable sensor devices are a growing industry field. Smart gadgets like fitness trackers monitor sport exercises and daily movements of the wearer. Sensor data is wirelessly transferred to smartphones and processed by sports and fitness apps. Depending on the application, a single sensor might not be sufficient. In ubiquitous computing, sport science and medicinal appliances, complex scenarios require multiple distributed sensors. However proper sensor data fusion is not possible without precise time synchronization among all units. The common wireless technology of wearables is Bluetooth Low Energy (BLE), which is integrated in every nowadays smartphone. But the architecture of BLE makes implementation of accurate time synchronization nearly impossible due to nondeterministic delays. In this paper we present two principles to obtain precise synchronization hardware time stamps, based on the power consumption of BLE System on Chips (SoC). Our concept exceeds the best-known accuracy for BLE and is more energy-efficient compared to related work. Using a common BLE SoC, we achieved a standard deviation of 0.9 microseconds.

Development of a quasi time stretch technology for indoor positioning system based on pulse modulated ultra high frequency radio

Data acquisition of high speed signal is a major challenge in developing a high accuracy Indoor Position System (IPS) based on Ultra High Frequency (UHF) Radio. The proposed technology is developed to achieve a high time resolution for Time Difference of Arrival (TDOA) estimation in a hyperbolic position fix IPS. This device works as the interface component embedded in the distributed sensor nodes of the IPS and is dedicated to convert the incoming UHF signal down to low frequency typically less than 1 MHz that enables the use of conventional relative low speed ADC product. The circuit model is simulated in Advance Design System (ADS) environment. Performance of the Wilkinson power combiner is evaluated using High Frequency Structural Simulator (HFSS).

A Wearable Flexible Sensor Network Platform for the Analysis of Different Sport Movements

In elite sports real-time feedback of biomechanical parameters is indispensable to achieve performance enhancement. Wearables including embedded data analysis are a suitable tool for online monitoring of movement parameters and might enhance the quality of training significantly. However, due to limited compute capacities for complex data processing on the sensor device itself, analysis can typically only be done afterwards using high-performance tools. This lack of immediate feedback may lead to slower training progress. We present a flexible, wearable system for the analysis of different sports movement including online-monitoring. It includes a modular, platform-based framework with a sensor node, an embedded software stack, Bluetooth Low Energy communication and an Android application. Data is analyzed on the sensor itself via embedded real-time algorithms. Results indicate that the device provides reliable and accurate measurements of movement parameters. In combination with adaptable algorithms and the BLE transmission, it offers solutions for real-time monitoring of athletic performance.

IMU-based determination of fatigue during long sprint

Stride parameters represent basic and useful information on track and field sprint performance. Contact mats or opto-electronic systems allow precise and unobtrusive measurements of those parameters, but their use is limited in space. Hence, there is a lack of research regarding the changes of temporal parameters throughout the competition distance (especially for long sprint), e.g. as a result of fatigue. Wearables, respectively inertial measurement units (IMUs), are not bound to limitations in space and therefore offer challenging opportunities for in-field diagnosis. This paper presents a wearable device for detecting and monitoring stance durations and step frequencies during sprinting. An application in (repetitive) long sprints is presented that analyzes changes of the temporal structure of performance parameters as a result of fatigue and level of expertise. Results indicate that the device provides reliable and accurate measurements of temporal parameters during sprinting and offers a deeper insight to movement characteristics of long sprint.

A Wearable Inertial Sensor Unit for Jump Diagnosis in Multiple Athletes

Flight and stance duration during jumping represent basic and very useful information for track and field coaches, and empirical evidence has been given that these parameters correlate strongly with elite performance (Hunter, 2004; Li et al. 2010; Slawinski et al. 2010). In highly dynamical sports such as track and field, athletes must be able to generate high forces within a very short time and in an appropriate manner. Consequently, reactive strength training including multiple jumps or drop jumps from different heights is very important for such athletes (Kale et al., 2009, Markovic et al., 2007). Objective feedback on performance is crucial to ensure a high quality of such a training as intrinsic information is merely available to the athlete due to the high movement velocities. From a trainer’s perspective, on the other hand, the quality of performance cannot be assessed precisely enough by pure observation. For the diagnosis of jumping performance in field-based conditions, several devices have been established in the last years. Contact mats or optoelectrical systems like Optojump® allow a precise and unobtrusive measureing of temporal parameters, but limitations must be stated according the operational area as well as group or ubiquitous monitoring. More recently, the availability of miniature solid-state inertial measurement units (IMUs) offers large opportunities to overcome these restrictions, and therefore open a new perspective for in-field diagnosis. Combined with wireless data transmission, IMUs can be used to provide athletes and coaches with fast and accurate performance measurements to improve athletic development and elite performance. Additionally, IMUs merely affect athletes during performance due to their small size and weight. IMUs have already been used to detect kinematic parameters in track and field applications. High correlations could be shown between IMUs and reference measurements (force platforms and optometric systems) for flight time and jump height during counter-movement-jumps (Picerno et al. 2011; r=.87) and for reactive strength index during drop jumps (Patterson and Caulfield, 2010; r=.98). Reactive strength index, for example, can be used for several purposes for the optimization of plyometric training or for injury prevention (Mc Clymont, 2003). It has also been applied as a tool to judge athletes’ recovery state (Horita et al. 1999; Toumi et al. 2006). Bergamini et al. (2012) reported mean differences of .005 seconds between IMU and highfrequency video or dynamometry for stance and stride durations during sprinting. Lower correlations between force and acceleration peaks for drop jumps (r=.70) and countermovement jumps (r=.55-59) were found if only a three-axis accelerometer data were considered (Tran et al. 2010). The aim of the recent study was the development and validation of an inertial sensor based device for detecting explosive jump events in elite athletes. Additionally, an ubiquitous group monitoring should be supported to use the device during training sessions with multiple athletes.