EN 
DE Usage scenarios for the acquisition and evaluation of biosignals
Author: Volkhard Klinger, FHDW Hannover, Embedded Systems Department
Contribution – Embedded Software Engineering Congress 2015
Summary
Modular and mobile systems are essential for the medium- and long-term monitoring of biosignals. The realization of a modular platform system for acquiring and analyzing biosignals, and its integration with sensors and actuators based on the Internet of Things (IoT) principles, enables new application scenarios. This paper presents two application examples and illustrates the corresponding adaptations to the platform-based concept. The application-specific integration of additional devices or sensors expands the platform's scope and flexibility. When external IoT systems are used, requirements for these systems arise, which are summarized below. It becomes clear that the advantages of the platform approach can only be maintained if these constraints are met.
Introduction
The acquisition and processing of electrical biosignals, such as electroencephalograms (EEG), electromyograms (EMG), and electroneurograms (ENG), enables a wide range of applications in research, medical diagnostics, and rehabilitation. EMG and ENG, in particular, are used to obtain information from the musculoskeletal system, including peripheral sensory information and data related to movement control. These signals and their analysis can then be used by physicians, medical personnel, and even patients or athletes for diagnosis or to monitor specific parameters. The spectrum of applications ranges, for example, from the therapeutic measurement of nerve conduction velocity during rehabilitation to complex prosthesis monitoring.
The requirements for a system for recording and evaluating biosignals are summarized in the following list.
Data acquisition and stimulation
The EEG, EMG or ENG data must be recorded and analyzed taking into account the signal characteristics (Table 1, PDF) are digitized. In certain applications, stimulation of nerve impulses is also necessary, for example when measuring nerve conduction velocity.
Data processing
The acquired data must be improved with regard to signal conditioning. Depending on the application, various filters and amplifiers are required for this purpose. Additional measures may also be necessary to handle asynchronous signals and non-equidistant samples. Furthermore, data processing plays a central role in data identification and correlation. While identification is necessary to assign groups of specific action potentials from the ENG [9] and to recognize control signals and corresponding feedback signals, data correlation aids in sensor fusion. Based on artificial intelligence concepts (machine learning), complex signal sequences are recognized and correlated with biological feedback mechanisms and additional external sensors. The identification method and a corresponding verification approach were introduced in [7] and [8] based on the results in [2], [3], and [4].
Data archiving
The recorded data must be stored locally if there is no direct transfer to a host system. Additional storage resources are also required for the identification process. The learned data is stored there for the operational phase [7].
Data exchange / Connectivity
The data must be transferred to a host system for evaluation or monitoring.
User Interface
The user interface allows the selection of different functions. At the same time, the collected and processed data is displayed.
configuration
The system functions can be configured for different use cases.
Use cases
Of the many possible applications, two are selected here that pursue very different goals. After a brief description of the medical circumstances, which are only addressed very abstractly here, the system architecture suitable for the application can be presented in the next chapter.
Nerve conduction velocity monitoring
A reduction in nerve conduction velocity can have many causes, such as demyelination of the axonal fibers or axonal damage. Neurophysiological measurements are certainly possible and are performed by stimulation followed by a measurement of the response velocity at a specific distance from the stimulation site. However, these measurements are usually performed only rarely, once a week or even less frequently. Regular and more frequent measurements are necessary to assess nerve conduction velocity in relation to factors such as the time of day, temperature, fatigue level, and the timing of medication administration. The influence of these parameters plays a crucial role in better evaluating changes. The primary requirements for a measurement system can be summarized as follows:
- Sensor interface
- Analog frontend for signal conditioning
- Mobility of the measuring system
- Local storage
- Data exchange / Connectivity
Prosthesis control and management
Today's prostheses are more than simple static replacement components. The challenge remains in the coupling, i.e., the information transfer between the human and the prosthesis. The approach used here utilizes coupling via the action potentials of the peripheral nerves, i.e., the use of ENG [6]. Using ENG and the identification method, it becomes possible to evaluate actuator and sensor data and correlate them with movement. However, this requires measurement data with high resolution and good signal conditioning. Furthermore, the inclusion of additional sensors is necessary to perform movement attribution with greater certainty. These external sensors can be implemented as microelectromechanical systems (MEMS) or as cameras. The use of the different sensors is based on the operating state of the system, which can be divided into a learning phase (stationary application of the system) and an operating phase (mobile application of the system). The correlation of the different sensors enables the use of forward and backward kinematic methods [5].
The primary requirements for the prosthesis control system can also be summarized briefly:
- Special sensor interface
- Complex analog front end for signal conditioning
- Mobility of the measuring system
- Local storage
- Identification method (learning and operational phase)
- High-performance local data processing
- Data exchange / Connectivity
3 System Platform
The data acquisition and analysis system was designed and implemented as a platform. Based on this platform, a modular hardware and software architecture is available, offering sufficient flexibility and scalability to meet diverse requirements. The software platform is characterized by the use of the hardware-independent, dynamic software platform OSGI (Open System Gateway Initiative). Details regarding individual aspects, such as sensor type, analog front end, etc., are omitted from this paper (see [7], [8]).
In Figure 1 (see PDFThe system is shown for the first application, the measurement of nerve conduction velocity. The platform architecture is illustrated by the different modules. A new aspect is presented here through the connection to a smartphone, which is connected via Bluetooth. This connection will be examined in more detail in Chapter 4.
In Figure 2 (see PDFThe system for the second use case is outlined in Figure 1. This is still a preliminary design; a System-in-Package (SIP) variant does not yet exist. The platform differs significantly from the first use case. This is due to the change in the analog front end (EMG → ENG) and the altered requirements regarding identification and data correlation. Additionally, several interfaces are present, which on the one hand transfer data from the implanted SIP to the body-worn system, and on the other hand exchange data and events between external systems, represented here as a smart device and MEMS sensor, and the platform. The learning phase required for identification at the beginning of system use is not shown here.
Requirements for IoT components
The analysis of the various use cases demonstrates the need to integrate additional sensors into the data acquisition and analysis system. This will specifically improve connectivity (A4). The proliferation of many so-called smart devices, such as tablets and smartphones, provides a potential infrastructure that can be used as part of the platform, enhancing its flexibility and scalability. Furthermore, the availability of intelligent components within the framework of the Internet of Things (IoT, [1]) is increasing rapidly. In this context, intelligent sensors can be connected to the platform. This decentralized peripheral significantly expands the platform's range of applications.
However, there are some key aspects to consider:
- The core platform remains necessary because basic functionality is only possible efficiently and with high performance through close coupling of system components.
- The modular nature of hardware and software in platform design is more necessary than ever as a design criterion.
- The use of software platforms in particular, OSGI being explicitly mentioned here, is what enables efficient coupling of platform systems and IoT components.
- The smart device must function as a gateway, providing interfaces at OSI levels 1 and 2, to enable seamless pairing [10]. Bluetooth alone, as a widely used communication interface, is often insufficient in this regard.
- The service orientation of the interface is a crucial aspect of integrating IoT components. Without an interface with a defined Quality of Service (QoS) characteristic, efficient coupling at the application or service level is hardly possible.
- Security aspects are another key consideration when connecting IoT systems. Without secure data transmission and a security infrastructure that, for example, operates on a certificate-based system, the use of IoT concepts is not feasible for many applications.
5 Summary and Outlook
The use of monitoring platforms based on platform architectures allows the system to be flexibly adapted to different application scenarios. The additional integration of IoT systems further expands the range of applications and enables data correlation, thus facilitating sensor fusion and context recognition.
The essential functions of each data acquisition system are shown in Figure 3 (see PDFThe platform circle comprises the components for communication (C: Connectivity), data processing (P: Processing), local storage (M: Memory), and the actuator/sensor interface, including the necessary analog front end (A/S: Actor/Sensor). These four components must be designed accordingly for each use case. The core of the system, which provides specific interfaces for the hardware and software platform, is also a central element. Here, OSGi plays a crucial role, capable of providing services at the application level.
If the application scenario changes, the platform must be adapted. This process is shown in Figure 4 (see PDF) shown.
This example assumes that the four basic components must be designed for the second use case. This means replacing modules to, for example, adapt the connection of the analog front end to the signal characteristics and sensor type. Furthermore, a more powerful processing unit and more local memory are required. The software platform is also modified and provides the corresponding services. These changes must then also accommodate the integration of IoT components. The requirements outlined in Chapter 4 then apply to these components. If these requirements are not met in the future, the efficient use of this concept will hardly be possible.
For future data acquisition and analysis systems, mapping the general platform cycle to the physical system level will be the crucial challenge (functionality -> device). The better both hardware and software components can be integrated into the platform, the sooner IoT systems will represent a viable and practical alternative to custom-developed modules with specific functionality.
literature
| [1] | A. Bassi, M. Bauer, M. Fiedler, T. Kramp, R. van Kranenburg, S. Lange, S. Meissner (Eds.). Enabling Things to Talk – Designing IoT solutions with the IoT Architectural Reference Model, 2013, Springer Berlin Heidelberg |
| [2] | Sebastian Bohlmann, Arne Klauke, Volkhard Klinger, and Helena Szczerbicka. Model synthesis using a multi-agent learning strategy. In The 23rd European Modeling & Simulation Symposium (Simulation in Industry), Rome, Italy, September 2011. |
| [3] | S. Bohlmann, A. Klauke, V. Klinger, and H. Szczerbicka. HPNS – a Hybrid Process Net Simulation Environment Executing Online Dynamic Models of Industrial Manufacturing Systems. In Proceedings of the 2009 Winter Simulation Conference MD Rossetti, RR Hill, B. Johansson, A. Dunkin, and RG Ingalls, eds., 2009. |
| [4] | S. Bohlmann, V. Klinger, and H. Szczerbicka. System Identification with Multi-Agent-based Evolutionary Computation Using a Local Optimization Kernel. In The Ninth International Conference on Machine Learning and Applications, pages 840–845, 2010. |
| [5] | JJ Craig. Introduction to Robotics: Mechanics and Control. Prentice Hall, New Jersey, USA, 2004 |
| [6] | Gold, C., Henze, D. and Koch, C., 2007, Using extracellular action potential recordings to constrain compartmental models, Journal of Computational Neuroscience, 23(1), pp. 39–58. |
| [7] | Arne Klauke, V. Klinger. Identification Of Motion-Based Action Potentials In Neural Bundles Using An Algorithm With Multi Agent Technology, Proceedings of the International Workshop on Innovative Simulation for Health Care, 2013, 978-88-97999-26-3; Backfrieder, Bruzzone, Frascio, Longo, Novak Eds. |
| [8] | Volkhard Klinger. Verification concept for an electroneurogram based prosthesis control. In Agostino Bruzzone, Marco Frascio, Vera Novak, Francesco Longo, Yuri Merkuryev, and Vera Novak, editors, 3rd International Workshop on Innovative Simulation for Health Care (IWISH 2014), 2014 |
| [9] | Neymotin, S., Lytton, W., Olypher, A. and Fenton, A.. 2011, Measuring the quality of neuronal identification in ensemble recordings. J Neurosci, 31(45):16, pp. 398–409. |
| [10] | T. Zachariah, N. Klugman, B. Campbell, J. Adkins, N. Jackson, P. Dutta. The Internet of Things Has a Gateway Problem, Proceedings of the 16th International Workshop on Mobile Computing Systems and Applications, 2015, Santa Fe, New Mexico, USA, ACM |
System and hardware development – our training & coaching
Do you want to bring yourself up to date with the latest technology?
Then find out more here MircoConsult offers training courses/seminars/workshops and individual coaching on the topic of Internet of Things/system and hardware development.
Training & coaching on the other topics in our portfolio can be found here. here.
Systems and Hardware Engineering – Expertise
Valuable expertise in the field of Internet of Things/system and hardware development is available. here Available for you to download free of charge.
You can find expertise on other topics in our portfolio here. here.
