From Capsil Wiki
Disclaimer: This Wiki is a work in progress and it has not yet been evaluated through a process of external review. No responsibility is accepted for errors or omissions or for any loss or damages which may result from the use of the information contained in the Wiki.
Sports and Wellbeing
Maintaining a fit and healthy lifestyle is essential through all phases of life, starting as a child, moving into adulthood, and, finally, old age. As an adult moves into the later years of life it is natural there will be a decline in health and ability to do certain activities. An important aspect of aging is enabling independent living and a high standard of living. One factor used to assess this is the ability to perform Activities of Daily Living (ADL). This ensures that someone is able to perform tasks such as washing and cooking that are essential for taking care of themselves. Using technology elderly people can be more involved in the management of their own health and wellbeing. It can help modify attitudes and behaviour to improve and maintain a healthy lifestyle through positive feedback and motivational instruction. Those suffering from chronic diseases such as morbid obesity, COPD, diabetes, and stroke, could use this technology to help make positive changes in their lives by changing diet, increasing exercise, and following rehabilitation treatments.
Even in later life, sports can be continued and enjoyed by the elderly. By keeping up physical fitness the risks of falling or sustaining injury can be reduced, however, staying motivated to exercise, measuring performance and ensuring good technique is vital. Monitoring and positive feedback can be used to measure improvements due to exercise, correct technique, and provide motivation to continue a program. Running, walking and rowing are examples of some of the sports that would benefit through monitoring.
During old age the chances of suffering from a chronic disease or falling increases. As a part of their recovery or treatment, physiotherapy is often prescribed to aid recovery, strengthen weakened muscles, and help rebuild confidence. While in hospital, following a program of physiotherapy is simple, as the frequency and content of the rehabilitation program is decided for the patient taking them through the exercises, offering encouragement, and helping correct technique. When a patient is discharged, maintain the course of physiotherapy can be challenging. Using similar techniques as those used to monitor sports; physiotherapy can also be monitored and managed, building a solid fitness base for recovery and increased fitness.
Sport Application Examples
Nutrition and Training Monitoring
By keeping a log of all food eaten and the activities, daily calories compared to energy expenditure can be calculated ensuring the optimum amount of calories is being consumed compared with the amount of daily exercise. This can also be used to customise a training program to help promote a balance between heavy and lighter training sessions. This will help build a good fitness base and promote fast recovery and improvement without injury.
Running / Walking
Running and walking are excellent for managing weight, improving endurance, and sustaining a healthy lifestyle. For the elderly, walking is particularly beneficial as it is low impact and requires no equipment, apart from either a pair of running or walking shoes. Measuring heart rate and step length during an exercise session can provide information relevant for estimating the distance travelled, speed, heart rate in beat per minute and calories burnt. These can be then incorporated to into a training program to build fitness and provide feedback to the user.
Rowing can provide a good all round fitness sport promoting muscle strengthening with flexibility and a cardiovascular work out. It is also relatively low impact and as such ideal for the elderly. Maintaining good posture and technique is vital for avoiding injury and maximising benefit. Traditionally a coach would observe from the bank suggesting changes in technique, however, this is not practical on a day to day basis. Wireless sensor networks (WSNs), for on-water rowing, and ambient sensors, for indoor rowing on a ergometer, could be used to monitor posture and suggest changes in technique.
There are many ways in which activity can be monitored depending on the type of data that is required. These could include; biomechanical, ambient, respiratory, and circulatory sensors. Home and mobile systems and WSNs can provide platforms for incorporating sensors and collecting data in a non-intrusive way.
There are several sensors that are suitable for measuring local body motions, such as those made by the limbs. To measure the angular change of joint angles, Electric Goniometers (EGs) can be used which provide a varying voltage output depending on the change in angle. While these sensors are cheap, they may prove awkward to attach by the user, especially if elderly, to be able to measure the relevant angles accurately.
Other sensors that have been used to measure motion are MEMS (MicroElectroMechanical System) accelerometers and gyroscopes, because of the decreasing size, cost, and energy requirements. Accelerometers have been used extensively, often placed on the limbs, to determine the limb motion. One of the problems of using accelerometers is that they measure both static acceleration with respect to gravity and dynamic acceleration. These can be hard to separate without additional information; however, dynamic acceleration can provide useful information regarding the impact of the body on other surfaces. Accelerometers can also be used to provide information on the bodies sway and general biomechanical motion.
MEMS gyroscopes have also been used to capture the motion of the body due to the issues involved in using accelerometers. By providing angular velocity information the rotation of the limbs can be found. Both commercial companies and the research community have used gyroscopes extensively to measure the body’s movements.
For applications such as running and walking force plates can be used to analyse the gait cycle. From this the distribution of the weight over the foot can be found and provide useful information regarding rehabilitation, the use of walking aids, and walking and running style. Conventionally force plates are incorporated into the floor limiting their range of use; however, developments in force sensors could provide for a portable version in the future that can be inserted into the shoe such as the Paro Tech insole product.
Detailed analysis of human motion can be captured using ambient or visual systems. These offer a wireless non-intrusive method for monitoring sports performance without interfering with the activity. There are many systems that are commercially on the market that use imaging systems, such as infrared cameras, to detect markers placed on the joints to track the motion of the user. One of the drawbacks to these types of system is the necessity of many cameras, with a range of only a couple of meters, and use of multiple markers which increase the complexity of the tracking and can suffer from occlusion.
If a high degree of detail is not required for motion analysis there are several alternative methods of motion capture that could be used. Passive Infra Red (PIR) sensors can be used to detect general motion in a room providing a high degree of privacy, but offering very limited motion resolution. Alternatively, a device called a blob sensor can be used to capture image data from a room and then convert the information into a binary image using a statistical model, background segmentation. Using the blob sensor, global pose can be derived; however, self occlusion can cause incorrect blob sensor results. Higher motion resolution can be found by using optical flow to capture motion within the blob.
Cardiovascular and Respiratory
Monitoring the cardiovascular system can provide valuable information related to general fitness and the health of the user. Electrocardiographs (ECG) measure the electrical activity of the heart through electrodes attached to the chest. ECG is often used to determine damage to the heart or the onset of disease. Implantable versions are also available for monitoring episodic events. Photoplethysmography (PPG) can also be used to monitor cardiac rhythm.
To measure respiration, spirometry can be used to measure the rate of the air transferred through the lungs, and the volume. This test can be used to determine the quality of the respiratory function for those suffering from Chronic Obstructive Pulmonary Disorder (COPD), asthma, emphysema, and cystic fibrosis, as well as athletes. Breath rate has also been measured using piezoresistive sensors, which provide a varying output as the force on them changes, attached across the chest.
Challenges and Issues
To develop systems that can be used to monitor and provide valuable feedback to the user regarding the performance of a given sport or activity, there are several challenges and issues that need to be addressed, many of which are common to other CAPSILs. These include: biosensor and platform development, power management, data modelling and inference, and user interfacing.
Biosensor and Platform Development
To capture the required information for sports activity monitoring, both a suitable sensors need to be developed and a wearable platform to interface with the sensors. Sensors that would be used for activity monitoring need to be small, light, non-intrusive, and easy to attach. Incorporating sensors into clothing is one method of attaching sensors. This presents challenges such as ensuring the position of the sensors does not change, making the sensors either detachable or robust to washing and designing garments that are specific to the user’s dimensions. If sensors are to be attached directly to the body they need to be made from biocompatible materials that will not irritate or harm the skin or user. In terms of platform development there are a number of WSN platforms available depending on the deployment and type of sensors used.
For any wireless system that will be deployed and expected to run autonomously for extended periods of time, power management becomes an important issue. Replacing and recharging batteries may not be possible or practical if sensors are implanted or hard to access. For implanted devices, induction coils could provide a means of recharging; however this would only be possible for implants near the surface of the skin. This has led to research into the field of energy harvesting to provide a solution which does not require external intervention. Current energy harvesting techniques include use of temperature difference, motion, and biological sources to generate power.
Data Modelling and Inference
In terms of data modelling and inference there are many different research areas including sensor placement, feature selection, data modelling and inference. Also, when multiple sensors will be used, synchronisation is an important consideration.
- Sensor Placement & Feature Selection: The question of which sensors to use, where, and how to deploy them is vital to acquire the relevant information for monitoring a specific activity or sport. Sensors could be deployed on the body or within the surrounding area. In terms of sports, ambient sensors provide a good solution as they are not attached to the body and as such would not obstruct the user in any way; however, they do require the user to stay within there area of operation. Wearable sensors on the other hand have to be attached to the user, but offer more freedom of movement. Another consideration regarding sensor placement is the amount of similar data required. By collecting similar data robust algorithms can be used to compensate for the failure of a sensor such that the quality of the end result does not suffer. This would however increase the number of sensors and the amount of data collected which in turn would have a direct impact on the processing time, energy expenditure and communicational costs. As well as sensor placement, the features extracted from the raw sensor data to monitor sports can be investigated. By finding the features that can best describe or distinguish features the overall amount of data can be reduced. As with the sensor placement, it may be beneficial to build in redundancy for important features and or remove those that are providing too much repletion.
- Data Modelling & Inference: To successfully model the sport being conducted, it is important to have detailed knowledge regarding the activity such that relevant data can be collected and correctly interpreted. Due to the complexity of many sporting or rehabilitation activities, this is especially important. Once data has been collected and features extracted, classification or clustering algorithms can be used for inference. Modelling algorithms can be used to extract underlying patterns from the data observation; however, the choice of algorithm is dependent on the type of data, how the data will be processed, and the amount of data. Popular algorithms include Bayesian Networks that assume prior knowledge, and Conditional Random Fields (CRFs) and Hidden Markov Model (HMM), which are temporal. The latter two are able to model multi-stage activities that evolve over time. Classification algorithms for inference have traditionally been centralised and applied after the data has been collected and combined from all sources. With the increasing computational power of sensor nodes and the need for real-time evaluation and feedback the trend is now towards distributed systems. This removes the need for a centralised node and can be more easily scaled.
Interfacing with the User
A crucial challenge for sports monitoring, is making the system as easy to use as possible. This especially key if the potential users are the elderly, who may not be as technologically inclined. Any monitoring system would need to have the following attributes to make them practical for deployment: self-managing, self-healing, autonomic, and be contextually aware.