Technological foundations and current status of a modified, low-risk form of competitive boxing (Box'Tag^®)

Abstract

Boxing-related activities are commonly used for fitness enhancement, but for many people fear of injury precludes participation in competitive boxing. Over the past six years, technological developments have contributed to the emergence in Australia of a modified, low-risk form of competitive boxing called Box'Tag^®. The rules of Box'Tag^® prohibit impacts to the head and any impacts above a moderate level of force. Contestants wear instrumented equipment that, in combination with a dedicated software package, allows automated impact detection and real-time display of scores. Participatory action research methods are being used to guide iterative technology refinements, with success evaluated based on feedback obtained through constant, direct interaction with the technology end-users. In concert with technological advances, the popularity of Box'Tag^® is steadily increasing, with programmes now established in three Australian states and about to be initiated elsewhere. Box'Tag^® is attracting demographically diverse participation and injury rates are low. Physiological and perceptual data recorded during contests show Box'Tag^® to be a highly intensive activity and an excellent vehicle for building physical fitness. There is scope for further technological improvements, but the Box'Tag^® initiative is already demonstrating potential to add a new dimension to boxing and has been adopted by Australian boxing authorities.

Keywords:

• Box'tag
• modified sports
• boxing
• scoring
• interactive textiles
• wearable sensors

1 Introduction

In Australia, boxing-related exercises and activities are widely used by the community for fitness development (Standing Committee on Recreation and Sport, 2011). By contrast, the number of people participating in formal boxing competition is small. The low rate of transition from training to competitive involvement is likely due to concerns relating to risk of injury. The concerns appear to have some foundation. Unterharnscheidt (1995) suggested that the repeated head impacts associated with boxing can lead to significant, permanent brain damage, causing loss of neurons, cerebral atrophy and functional deficits that may eventually manifest in clinical symptoms such as traumatic Parkinsonism. It was argued that the pathology could occur in both amateur and professional boxers. McCrory, Zazryn, and Cameron (2007) subsequently concurred with this view, although they did note that an increasing medical presence in boxing, along with other changes to the sport, could be expected to reduce the future incidence of chronic traumatic encephalopathy. Tanriverdi et al. (2010) recently reported that anti-hypothalamic and anti-pituitary antibodies were each raised in more than a third of 61 amateur boxers studied, but in none of 60 age-matched controls. They postulated that the elevated antibodies were due to head trauma and could well be the mechanism of the disturbed pituitary function sometimes observed in boxers. The dangers of repetitive head trauma have been further highlighted by Amen et al. (2011) who conducted high-resolution brain spectral imaging, quantitative electro-encephalography (qEEG), and various neuro-psychological tests on 100 current and former American football players and compared the results with reference data obtained from the general population. Overall, the footballers exhibited significantly decreased brain perfusion, more qEEG abnormalities and relatively poor performance on the neuro-psychological tests. The perfusion changes were not conclusively different between players who had experienced a high number of loss-of-consciousness episodes and those who had experienced none, perhaps indicating that they were due more to repetitive sub-concussive impacts than to a few major traumas.

Pearn (1998) has argued that, considering the danger of chronic brain injury, young people should be banned from participating in boxing until they are 16 years of age and capable of understanding the risks and making an informed decision.

In view of the above, there is a strong case for development of a modified form of competitive boxing that minimises the risk of injury to contestants and encourages broader participation. Six years ago, a Sydney-based boxing and fitness company decided to take up this challenge and introduced an activity named Box'Tag^®. The subsequent evolution of Box'Tag^® has been driven by the design, production and implementation of specialised technologies, and we have been among the contributors to that process.

The purpose of this paper is to report on the outcomes achieved to date.

2 Experimental

We have used the participatory action research methodology outlined by Kemmis and McTaggart (2005). This methodology involves repeated cycles each characterised by stages described by the words: Reflect, Plan, Act, Observe. It is designed for implementation in ‘real-world’ situations and involves the direct participation of the people for whom solutions are being sought. It is seen as an ideal methodology when rapid and/or holistic change is required. Our focus is on developing technology that can play a major role in making Box'Tag^® safe, enjoyable for participants, affordable, attractive to spectators and a good vehicle for skill and fitness acquisition. Success is being measured not only in terms of the progression and performance of the technology, but also on the basis of observations with respect to the uptake of Box'Tag^®, the incidence of injuries, the physiological intensity of the modified sport and a range of other factors connected with the practical implementation of Box'Tag^® programmes.

In keeping with the principles of participatory action research (Kemmis & McTaggart, 2005), we see the whole field of practice of Box'Tag^® as part of our experimental environment. We are therefore continually collecting both qualitative and quantitative data, and as the field of practice expands, both the volume and richness of the data are increased. Box'Tag^® competitions and training sessions provide opportunities for direct observation that can then be used to help guide subsequent processes of reflection, planning and action aimed at improvement. Constant dialogue with organisations, participants, coaches and club administrators involved in Box'Tag^® programmes is considered essential to the efficacy of the reflective, planning and action phases of each iterative participatory action research cycle. Documentation of all phase outcomes is also critical. Written project reports are prepared on an approximately quarterly basis and distributed to interested parties for feedback. A trial is currently being conducted of a software package (CutOrange, developed by Cordelta) aimed at stimulating more spontaneous interaction and documentation through what is effectively a customised social networking model.

All of the experimentation is being carried with a clear focus on the overall end-goal – development of Box'Tag^® as a viable, safe, challenging, attractive and enjoyable sport.

3 Results

3.1 Overview of current Box'Tag^® scoring technology

An automated impact sensing system, developed through collaboration between the Commonwealth Scientific and Industrial Research Organisation (Australia's primary government-supported research agency) and the Australian Sports Commission, is used for scoring Box'Tag^® contests. Because the rules of Box'Tag^® prohibit impacts to the head (thereby reducing the probability of brain trauma to almost zero), it has been necessary to redefine the target zone to include just the torso and small areas on the upper arms. The latter are not legitimate scoring targets for conventional boxing, but have been added in the Box'Tag^® context so that the total area that has to be defended is not unduly compromised.

Box'Tag^® contestants wear T-shirt style instrumented vests and standard gloves with patches of conductive material affixed to their surfaces. The vests incorporate stripes of silver-coated nylon yarn through which a low-level electric current can be run. When two parallel stripes are contacted by the conductive portion of a glove, a change in electrical resistance occurs. Small transceivers connected to the vest detect the change in resistance and send the information via Bluetooth to a ringside computer where a dedicated software package applies customised algorithms to determine whether a score should be registered. Scores are displayed in real time. The vests and glove patches worn by contestants can be seen in Figure 1, which also highlights the fact that despite the veto on head impacts, contestants are required to wear head guards and mouth guards to ensure that they are well protected from any inadvertent contacts. (In addition, it is mandatory for male and female contestants to wear groin guards and chest guards, respectively).

Figure 1 Box'Tag contest showing instrumented vests and gloves. (Photograph and © by Kris Arnold of Kris Arnold Photography, 2012, reproduced with kind permission.)

3.1.1 Instrumented vests

The design of the vests has been described in more detail by Helmer et al. (2010). A sensor fabric is sewn on to the surface of a nylon base garment in patterns that delineate the target areas. The fabric – which is formed on a Jumberca Mini Jacquard double knit machine model 4TJ (28 gauge, 30 inch diameter, 48 feed) – consists of three polyester layers, and the stripes of silver-coated nylon yarn are knitted into the top layer, forming electrodes spaced 40 mm apart. Scouring, stenter drying and heat setting of the fabric is carried out to stabilise it. A hydrophobic finish (Oleophobol 7713) is applied to reduce the possibility of fabric wetting causing shorting between electrodes and to increase the electrical resistance of such shorting. Laboratory testing of the vests has shown that while wetting does cause changes in resistance, its effect can be clearly differentiated from that of impacts.

The functionality of the vests has proved robust in the face of repeated washing over a two-year period during which some of the vests have undergone up to 20 wash cycles.

Figure 2 depicts various aspects of the vest production process.

Figure 2 Construction of Box'Tag vests: (A) knitting of sensor fabric on Jumberca Mini Jacquard double knit machine; (B) monitoring of yarn tension on knit machine to optimise fabric quality; (C) pattern layouts for producing vests of different sizes while retaining required positioning of sensor electrodes; (D) electronic connection of sensor electrodes via fabric running horizontally across torso region of vest and flexible cable running from shoulder target regions.

3.1.2 Conductive glove patches

Two different types of conductive glove patch have been used. Initially rubberised silicon strips (Jehbco, Sydney, Australia) were glued to glove surfaces. Because the maximum available strip width was 45 mm, it was necessary to conjoin two strips to fully cover the scoring region of the glove. The patches were found to retain conductivity over a period of nearly two years of regular deployment, but the process of affixing them to gloves was technically difficult and highly time-consuming, and they were susceptible to detachment during competition. Two-layered knit patches, with the under layer incorporating silver-coated yarn, are now being used (see Figure 3). These patches can be glued to gloves much more easily and seldom come adrift, but initial prototypes showed quite rapid and unpredictable losses of conductivity. Later versions incorporating the same silver-coated thread as the vests are more durable. The extent of the improvement is still being assessed. At the same time, experimentation is being carried out with a new, wider (90 mm) rubberised strip, but the challenge of securing the strip to gloves in a reliable and time-efficient way persists.

Figure 3 Two-layered knit conductive patch attached to surface of boxing glove (also shown is the transceiver used to record signals from vest sensor electrodes and transmit them via Bluetooth to a ringside computer). (Photograph and © by Kris Arnold of Kris Arnold Photography, 2012, reproduced with kind permission.)

3.1.3 Software

The general characteristics of the software package (Spartan; developed at the Australian Institute of Sport) have been outlined by Bruch et al. (2011). The software reads the electrical resistance data, streamed in real time as 8 bit values at 250 Hz, and determines whether they meet set criteria for registration of a valid impact. The current iteration of Spartan (version 4.9) uses a 10-s moving average of the vest electrical resistance data to nullify the effects of any gradual changes in vest condition (e.g. those that may result from athlete sweating). A score is recorded when resistance falls to less than 80% of the moving average for a period of between 10 and 500 ms, with this period considered to define the range of glove contact time consistent with a legitimate punch, as opposed to a glancing blow or a push. The Spartan software is capable of capturing video images at a rate of 25 frames per second from two elevated, fixed cameras (Basler BIP-640C) usually located at right angles to each other. It allows synchronised frame-by-frame replay of the two video streams integrated with the sensor data. This not only provides a valuable tool for analysis of athlete performance, but also enables comprehensive post-contest evaluation of the accuracy of the scoring system.

3.2 Accuracy of scoring

Two years ago, Bruch et al. (2011) found that the scoring system was detecting ∼90% of all impacts that on the basis of the frame-by-frame video examination appeared to be valid. More than a third of the missed scores were associated with impacts to the shoulder target areas, which are relatively small and include only two sensor stripes. Variability in the time taken for transmission of Bluetooth data packages was another source of the error. The latter problem has since been addressed by modifying transceiver firmware so that it numbers the packages, and by adjusting the software so that it considers consecutive packages as having arrived at intervals of 4 ms (given the data rate of 250 Hz). Also, action is presently in progress to increase the number of sensor stripes in the shoulder regions of the vests. Consequently, the incidence of missed impacts should soon be considerably decreased from an already quite low level. False positive scores are rare and, in the study of Bruch et al. (2011), occurred at an overall rate of only 0.15 scores per minute per contestant.

3.3 Uptake of Box'Tag^®

A Box'Tag^® programme has been operating since 2006 at the Sydney club that initiated the concept. As part of this programme, regular competitions have been held and in total there have hundreds of contests, usually consisting of three rounds of 1–2 min separated by 1-min rest intervals. The competitions have attracted participants of diverse age, socio-economic background and cultural heritage. Many of the contestants have been people who would never consider participation in conventional boxing because of concerns about the potential for injury. Almost a quarter have been female.

Substantial enhancements to the scoring technology over the past 2.5 years have been associated with a considerable expansion of Box'Tag^®, and programmes are now established in Melbourne and Canberra. In 2011, a pilot programme was carried out at a school in a disadvantaged area of Brisbane, with the support of the Queensland Academy of Sport Centre of Excellence for Applied Sport Science Research and Griffith University. It attracted high levels of voluntary participation and adherence and led to significant gains in the physical fitness of the students. The school reported that several of the participants showed dramatic improvement in overall classroom and school ground behaviour.

Box'Tag^® was successfully included in the 2012 Victorian Police & Emergency Services Games. This marked the return of competitive boxing to the Games after an absence of 23 years, following a 1989 decision to discontinue a conventional boxing competition. The organisers of the Games have already confirmed their wish to retain Box'Tag^® and have initiated discussions aimed at its incorporation into the next Australasian Police & Emergency Services Games.

Boxing Australia (the organisation that controls Olympic boxing in Australia) has officially adopted Box'Tag^® as its national ‘community participation programme’ and is committed to supporting its wider uptake. It has created strong impetus for establishment of new centres in New South Wales and Western Australia. There is interest also from elsewhere in Australia, and there have been some enquiries from overseas. A Box'Tag^® association linked to Boxing Australia was recently formed, and already has ∼200 members.

3.4 Safety of Box'Tag^®

The Box'Tag^® scoring technology rewards light, rapid punching with a view to contributing to participant safety. Injury rates in Box'Tag^® have been very low compared to those associated with most other sports. Accidental head impacts have occasionally occurred, but have never caused a concussion. The most common injury has been forearm bruising and abrasion caused by clashing with the arms of an opponent, but even this occurs quite seldom. Although the rules of Box'Tag^® entail deductions of points for impacts above a moderate level of force, there have been two reported cases of rib bruising. In addition, there has been one instance in which a contestant suffered a lower back muscle tear during a bout and was unable to continue, but overall it is evident that the aim of providing Box'Tag^® participants with a relatively safe competitive sporting environment is being met.

Box'Tag^® has been fully endorsed by a leading Australian paediatrician, Professor John Pearn, who has opposed conventional boxing on medical grounds (Pearn, 1998).

3.5 Physical intensity of Box'Tag^®

Measurement of heart rates and blood lactate concentrations during Box'Tag^® contests has shown that the modified sport is very physically demanding. Approximately half of the contestants studied by Lazarus (2011) recorded peak heart rates above 90% of age-predicted maximum and about a third reached blood lactate levels of greater than 10 mmol.L^− 1. Fifteen of 18 contestants who provided subjective ratings of their perceived exertion indicated that they experienced the required effort as being in the range between hard and maximal (Lazarus, 2011). In general, the figures are quite similar to those reported for conventional boxing competition (Arsenau, Mekary, & Léger, 2011; Ghosh, 2010; Khanna & Manna, 2006). The high physical intensity of Box'Tag^® may be due to a much diminished fear of injury encouraging almost constant engagement with the opponent. It is clear that preparation for Box'Tag^® competition requires considerable attention to the building of specific physical fitness, and that Box'Tag^® itself is an excellent activity for fitness development.

3.6 Cost of technology

Since the end of 2009, the cost of producing the scoring technology required for Box'Tag^® has been reduced by an order of magnitude. Moving to the present T-shirt vest style has been partly responsible for the reduction because earlier vest prototypes, which incorporated 16–22 polyvinylidene fluoride (PVDF) piezoelectric sensors (Hahn et al. 2010), were difficult to manufacture. Also, each contestant is now required to wear only one transceiver, which is located in a pocket on the back of vest, whereas previously there was a need to wear four (one in the vest, one in each glove, and one in the head guard). Essentially, cost decrease has been accomplished by greatly lessening system complexity, which – although reducing the amount of information provided by the system – has also substantially increased system robustness and reliability.

4 Discussion

Box'Tag^®, inseparably underpinned by innovative technology, is gaining significant momentum and is poised to add an entirely new dimension to the sport of boxing, initially in Australia but eventually also on the world stage. The low risk of injury favours longevity of participation. In some states of Australia, engagement in conventional boxing competition cannot legally commence before the age of 14 years, and under international rules participation in Olympic boxing is not permissible beyond the age of 34 years. By contrast, Box'Tag^® can be an almost lifelong pursuit. In addition, the relative safety of Box'Tag^® allows a high frequency of involvement in competition. Indeed, round-robin competition formats have proven practicable, whereas the nature of conventional boxing effectively requires the use of elimination formats.

The objectivity of the scoring system in Box'Tag^® is well regarded by contestants, and the scoring technology offers a number of advantages, including the ability to influence the strategies adopted by contestants. For example, the introduction of shoulder scoring regions to compensate for removal of the head from the target region has helped to ensure that the style of boxing typically used (in terms of such factors as ring movement, footwork, selection of distance from opponent) is reasonably similar to that associated with conventional boxing. Furthermore, the technology has potential for use as a tool for teaching of skills, since rewards for particular actions can be manipulated, as might occur if the conductive material on one glove of an athlete was either removed or decreased in size.

The Box'Tag^® scoring technology has come through an extensive proof-of-principle phase and has now reached the point of being an advanced prototype. To take the next step, there will have to be a focus on product development, with consideration given to the logistics of manufacturing, distributing and servicing system components in a commercial context. Funding of product development processes will be a substantial challenge, but discussions with manufacturing companies have been initiated and, interestingly, have raised a likelihood that there might need to be a further progression in the direction of simplicity to minimise both costs and potential for occurrence of technical problems.

In concert with an emphasis on product development, it is hoped that it will possible to continue R&D activities to enable future refinements to the product. There are several issues that merit further attention. At present, it is possible to generate a score for the opponent through contact between one's own glove and vest. The possibility has been minimised by carefully controlling the size and placement of the conductive patches that define the scoring region of the glove, but ideally should be reduced to zero. While the present version of the scoring system is quite sensitive in detecting impacts, it provides no indication of impact forces, which means that imposition of penalties for excessively forceful punches is at the discretion of the referee. Ability to measure or at least categorise impact forces would be a valuable enhancement of the system. Automated detection of head impacts is already possible, but as a pre-requisite to automated imposition of penalties in the Box'Tag^® setting, any head impacts occurring through self-contact would have to be definitively excluded. Otherwise, self-contact to the head would disadvantage the opponent. Currently, the identification of prohibited head impacts, like the detection of overly forceful impacts, is a referee responsibility. It can be argued that there are positives in retaining some arbitration role for the referee, but it does compromise the concept of total scoring objectivity.

The participatory action research approach has highlighted the need to be always on the lookout for ways to further augment the safety of Box'Tag^® and has led, for instance, to current experimentation aimed at producing specialised Box'Tag^® gloves capable of substantially absorbing impact forces. Through wide consultation and discussion, several different ideas have emerged as to how this might best be achieved, and two of the ideas have already progressed to the development and testing of prototypes. At the same time, experimentation is under way with a view to enabling routine measurement of impact forces during Box'Tag^® contests so that penalties for excessive force can be objectively and reliably imposed.

There is scope for applying the scoring technology used in Box'Tag^® in a range of other settings. The most obvious of these is Olympic boxing, which has been beset by major scoring controversies over many years (Hahn et al. 2010), and which despite continual refinements to scoring methodology since 1988 still experienced some notable problems during the 2012 London Olympic Games. A comparatively early version of the automated impact sensing system was demonstrated to the Referees and Judges Commission of the International Boxing Association (AIBA) in 2007 and stimulated some interest, but it was noted that the system required further work, and the priority of the Commission was on the continuing refinement of its existing scoring system. In view of the technical progress that we have made since that time, further consideration of the feasibility of adopting the automated system might now be justified. However, AIBA has decided that head guards will no longer be used by senior male participants in Olympic boxing, meaning that instrumentation of this equipment could not be used to detect head impacts. Alternative methods for discerning the impacts probably could be devised, but at this stage we consider it best to remain concentrated on the Box'Tag^® application. Representatives of other combat sports, particularly Muay Thai and various forms of karate, have enquired about the potential for adapting the technology for their use, and we might eventually need to pursue these avenues to enable realisation of the commercial returns required to support further work on the technology.

The development of Box'Tag^®, with its integral technological component, has been a remarkably collaborative endeavour, with more than 30 organisations having made some contribution along the way. The organisations have included dedicated research agencies, institutes and academies of sport, a national sporting federation, universities, technology companies, commercial fitness centres and community clubs. Their diversity has provided us with a comprehensive range of perspectives and has allowed synergistic progression of ideas. It has been and continues to be a fundamental strength of the project.

5 Conclusions

The technology being used in Box'Tag^® is playing a key role in allowing the modified sport to achieve its original aims. Criteria identified as markers of overall project success are being met. While there is room for further enhancement of the technology, the current iteration is performing at a level that is stimulating accelerating uptake of Box'Tag^® and broadening its demographic reach. The technology has features that are likely exerting a positive influence on safety of participants and promoting dynamic physical activity patterns. Initiation of a product development phase is now required. The advent of Box'Tag^®, enabled by technology, could well prove to have transformational effects on the sport of boxing.

Acknowledgements

The work summarised in this paper has been supported by grants from the Australian Institute of Sport, the joint research fund of CSIRO and the Australian Sports Commission, and the Queensland Academy of Sport Centre of Excellence for Applied Sport Science Research. The authors also wish to acknowledge that they have built upon work initiated by the Cooperative Research Centre for Microtechnology, in which Griffith University, RMIT University and Swinburne University were the primary academic institutions. Inputs from several Australian technology companies have also been important, particularly PWP Designs, Catapult Sports and Hydrix. Two honours students from Victoria University – Henner Bruch and Brendan Lazarus – have made notable contributions to the collection and analysis of data. Special thanks are due to the staff and members of the Strongarm Boxing & Fitness Club in Sydney, where much of the field-testing of the Box'Tag^® scoring technology has been performed.