Integrating physiological monitoring systems in military aviation: a brief narrative review of its importance, opportunities, and risks

Abstract

Military pilots risk their lives during training and operations. Advancements in aerospace engineering, flight profiles, and mission demands may require the pilot to test the safe limits of their physiology. Monitoring pilot physiology (e.g. heart rate, oximetry, and respiration) inflight is in consideration by several nations to inform pilots of reduced performance capacity and guide future developments in aircraft and life-support system design. Numerous challenges, however, prevent the immediate operationalisation of physiological monitoring sensors, particularly their unreliability in the aerospace environment and incompatibility with pilot clothing and protective equipment. Human performance and behaviour are also highly variable and measuring these in controlled laboratory settings do not mirror the real-world conditions pilots must endure. Misleading or erroneous predictive models are unacceptable as these could compromise mission success and lose operator trust. This narrative review provides an overview of considerations for integrating physiological monitoring systems within the military aviation environment.

Practitioner summary: Advancements in military technology can conflictingly enhance and compromise pilot safety and performance. We summarise some of the opportunities, limitations, and risks of integrating physiological monitoring systems within military aviation. Our intent is to catalyse further research and technological development.

Abbreviations: AGS: anti-gravity suit; AGSM: anti-gravity straining manoeuvre; A-LOC: almost loss of consciousness; CBF: cerebral blood flow; ECG: electrocardiogram; EEG: electroencephalogram; fNIRS: functional near-infrared spectroscopy; G-forces: gravitational forces; G-LOC: gravity-induced loss of consciousness; HR: heart rate; HRV: heart rate variability; LSS: life-support system; NATO: North Atlantic Treaty Organisation; PE: Physiological Episode; PCO₂: partial pressure of carbon dioxide; PO₂: partial pressure of oxygen; OBOGS: on board oxygen generating systems; SpO₂: peripheral blood haemoglobin-oxygen saturation; STANAG: North Atlantic Treaty Organisation Standardisation Agreement; UPE: Unexplained Physiological Episode; WBV: whole body vibration

Keywords:

• Safety
• human factors engineering
• hypoxia
• hyperoxia
• G-forces

1. Introduction

In military aviation, mission requirements and operational environments require the successful interaction of human and technological capabilities. The rapid advancement of military aircraft since the first World War now exposes pilots to physical and environmental extremes that can extend beyond what can be safely tolerated by their physiological limits. Failure to understand how newer generation aircraft can be incompatible with human physiology has increased the risk of inflight Physiological Episodes (PE) (Elliott and Schmitt 2019). A PE occurs when a pilot experiences symptoms or performance impairments beyond what is deemed acceptable. Whilst some PEs are minor and inconsequential, more severe incidences can lead to accidents, mission failure, loss of aircraft, and death (United States Department of Defence Inspector General 2021). Pilots are the most valuable components of human-rated aircraft; therefore, their inflight physiological responses should be monitored. The data can be used to enhance aerospace engineering and potentially integrate within the aircraft’s operational system to promote flight safety and efficiency. Reliable biomonitoring systems that continuously measure pilot physiology to predict and classify cognitive and functional states inflight are only in development and not yet fully operationalised (Brunyé et al. 2021; Pongsakornsathien et al. 2019).

The aim of this narrative review is to provide a brief overview of the physiological challenges military pilots may experience and to discuss possible opportunities, limitations, and risks for integrating physiological monitoring systems in military aviation. We intend for this review to share some of the critical challenges we currently experience in military aviation with a wider community with the hope of catalysing further research and development. Our primary focus is pilots of high-performance aircraft, but other aircrew and aircraft are included. To facilitate our discussion, we indiscriminately searched (without a specified inclusion and exclusion criteria) peer-reviewed journal databases, peer and non-peer-reviewed publicly available military and governmental reports, and media publications. Protected and unpublished military- and government-related research were not included.

2. Physiological stressors in military aviation

Military aircraft are diverse in technology and capability. These include fixed-wing high-performance fighter jets (e.g. F-35), large transport (e.g. Hercules) and reconnaissance (e.g. P-8 Poseidon, U-2) aircraft, as well as rotary-wing aircraft (e.g. SH-2 Seasprite and Chinook). The largest physiological threat to the pilot appears to occur in high-performance aircraft, which require the integration of wearable life-support systems (LSS). Early aircraft generations have progressed from maximising performance (e.g. speed, range, manoeuvrability) to integrating advancements in mission, avionics, and munition systems (e.g. stealth, surveillance, weapons) and can now operate as swing and multi-role fighters capable of warfare, reconnaissance, surveillance, and support. Whereas, due to less extreme flight profiles and reliance on LSS in large transport and reconnaissance aircraft (e.g. the continuous use of oxygen supply masks is not required due to pressurised cabins), the risk of a physiological threat is lower. Collectively, modern aircraft can now support network-centric missions via large-scale data management and communication with other aircraft. This highlights an opportunity for integrating human physiological data to promote flight safety and mission success. Although advances in automation and technology, including remotely controlled aircraft, may ultimately leave the human pilot redundant, they are currently an integral component of most military aircraft.

Initial challenges in aerospace engineering to protect the pilot required combatting hypobaria at high altitudes that can cause hypoxia (i.e. insufficient oxygen availability due to a low inspired partial pressure of oxygen [PO₂]) and decompression illness (West 2016). Pressurised cockpits and full pressure suits now allow aircraft to ascend to altitudes over 70,000 ft. However, cockpits can depressurise unexpectedly, and pressurised altitude equivalents can oscillate by ∼2000 ft during certain flight manoeuvres (e.g. avoiding missiles or aerial combat). Also, in some smaller, high-performance aircraft, cabin pressurisation is not sufficient to maintain the PO₂ at high altitudes, which necessitates the use of oxygen supply systems. Oxygen is either stored or generated and is delivered to the pilot at fractions between 21 and 100% when cabin pressurisation is above 10,000 ft. Hyperoxic gas (i.e. a PO₂ above the equivalent of 21% oxygen at sea-level atmospheric pressure) can cause hyperoxia and oscillations in oxygen fractions from on board oxygen generating systems (OBOGS) have also been reported (i.e. rapidly oscillating hyperoxic conditions). Pilots may also experience simultaneous changes in barometric pressure and oxygen supply (e.g. normobaric hyperoxia or hypobaric hyperoxia).

Accelerative (inertial) forces from manoeuvrable, high-performance aircraft increase gravitational-forces (G-forces). Force can be applied in two directions along the three main axes (i.e. ±G_x, G_y, and G_z) (Pollock, Hodkinson, et al. 2021), with + G_z force (i.e. cranial-caudal direction) having the most pronounced effect on human physiology. Anti-gravity suits (AGS), chest pressure garments, pressure breathing, and implementing the anti-gravity straining manoeuvre (AGSM) are now required for aircraft that generate >3 +G_z. Push-pull manoeuvres (i.e. −G_z followed immediately by +G_z) (Banks et al. 1995), poorly fitted AGSs, failure of AGSs to pressurise (Eiken and Grönkvist 2013) or poorly performed AGSMs (Tu et al. 2020) limit +G_z tolerance. Advances in AGS technology to mitigate push-pull effects have also integrated lower body negative pressure during −G_z to increase subsequent +G_z-tolerance (Xing et al. 2020).

Pilots may also experience other physiological stressors considered less threatening than overt aviation-related stressors, such as whole-body vibration, heat, hypohydration, and fatigue (e.g. sleep restriction, circadian desynchronisation). However, these can alter tolerance to another stressor, such as heat and hypohydration reducing G-tolerance (Nunneley and Stribley 1979), and increase the risk of losing situation awareness and causing spatial disorientation.

3. Functional and physiological responses to aviation-related stressors

Current understanding of aerospace physiology is derived almost exclusively from investigations of isolated stressors within controlled laboratory settings. However, there is an increasing number of publications reporting infight physiological responses in military aircrew, such as during attacking and defensive manoeuvres in high-performance aircraft (Hormeño-Holgado and Clemente-Suárez 2019) and helicopter rescue manoeuvres (Vicente-Rodríguez et al. 2020). An early fundamental and comprehensive synthesis of this topic can be referred to in the Bioastronautics Data Book (Parker and West 1973). Whilst a comprehensive overview of physiological responses to all aviation-related stressors is outside the scope of this review, the following paragraphs highlight some foundational physiological responses to the stressors previously stated.

3.1. Hypoxia

Hypoxia is a state of inadequate oxygen availability to maintain normal physiological function. The hallmark indicator of a hypoxic environment is hypoxaemia, a reduction in arterial PO₂, which is typically determined by reduced peripheral blood haemoglobin-oxygen saturation (S_pO₂) to <88–92%. Cognition, vision, and motor-performance are impaired, particularly under severe hypoxia (e.g. exposure to >20,000 ft PO₂ equivalent) (Shaw, Cabre, and Gant 2021). Hypoxia also reduces G-tolerance (Besch et al. 1994) and decreases AGSM performance due to impaired muscular performance (Millet et al. 2012). Hypoxaemia can also reduce cerebral tissue oxygenation (Ottestad, Kåsin, and Høiseth 2018; Phillips et al. 2009; Williams et al. 2019), despite an increase in cerebral blood flow (CBF) (Ainslie, Hoiland, and Bailey 2016; Hoiland et al. 2016). Ventilation increases in an attempt to elevate alveolar PO₂ and consequently decreases alveolar and arterial carbon dioxide, leading to hypocapnia (see below). Heart rate (HR) increases (Botek et al. 2015; Krejčí, Botek, and McKune 2018), HR variability (HRV) decreases (Botek et al. 2015; Krejčí, Botek, and McKune 2018) and electroencephalogram (EEG) power amplitudes are altered (Rice et al. 2019). These physiological responses appear greatest with hypobaric compared with normobaric hypoxia (Coppel et al. 2015).

3.2. Hyperoxia

Hyperoxia is a state of excess oxygen supply to tissues and organs.Breathing a higher PO₂ (at a barometric pressure of approximately sea-level) can improve cognition, particularly at inspired oxygen concentrations of ∼30–45% (Hayes, Temme, and Onge 2020). Hyperoxia saturates haemoglobin with oxygen (i.e. S_pO₂ > 98%) and increases dissolved oxygen within the blood (Cheng 2012). Pre-breathing 100% oxygen to denitrogenate the body prior to high-altitude exposure also reduces the risk of decompression sickness (Vann et al. 2011). Although hyperoxic breathing is required to maintain alveolar PO₂ and S_pO₂ at high altitudes (i.e. hypobaria), it can exert adverse physiological effects at low altitudes, despite concomitantly improving cognition (Damato et al. 2020). Hyperoxic breathing inflight also increases the risk of absorption and acceleration atelectasis (Dussault et al. 2016; Pollock, Gates, et al. 2021), producing cough, chest pain, and breathing difficulties. Cerebral blood flow subsequently declines to prevent over-oxygenating the brain (Lambertsen et al. 1953; Mattos et al. 2019). Ventilation initially decreases (Marczak and Pokorski 2004), then increases (Becker et al. 1996; Marczak and Pokorski 2004), which may lead to hypocapnia due to increased carbon-dioxide off-loading. Heart rate decreases and HRV increases (Bak et al. 2007; Gole et al. 2011; Lund et al. 1999), and EEG power amplitudes are altered (Damato et al. 2020; Kizuk et al. 2019). The functional and physiological effects of rapidly oscillating hyperoxic breathing remains uncertain, but to our knowledge are currently under investigation.

3.3. Hypocapnia and hypercapnia

Hypocapnia and hypercapnia are states of low and high arterial partial pressures of carbon dioxide (PCO₂), respectively. Hypocapnia results from hyperventilation, such as in response to hypoxia (Friend, Balanos, and Lucas 2019), prolonged hyperoxia (Becker et al. 1996; Marczak and Pokorski 2004), vibration (Lamb and Tenney 1966) or psychological stressors (e.g. anxiety) (Suess et al. 1980). Hypocapnia presents similarly to hypoxia (Shaw, Cabre, and Gant 2021), whereas hypercapnia presents with tachycardia, confusion, shortness of breath, and headaches. Hypercapnia results from inhibited ventilation and limited carbon dioxide off-loading, such as restricted breathing systems (e.g. mask air flow resistance) or due to an inability for full ventilation (e.g. atelectasis). Carbon dioxide has a strong influence on ventilation, CBF (Ogoh 2019), and haemoglobin affinity for oxygen (Stepanek et al. 2020). Arterial PCO₂ is the gold standard measurement, but it can also be inferred from the end-tidal partial pressure of carbon dioxide (McSwain et al. 2010), although this may not be reliable across all aerospace conditions (Shykoff et al. 2021). Hypocapnia reduces CBF, whereas hypercapnia increases CBF (Ito et al. 2003). Heart rate increases in hypocapnia (Rutherford, Clutton-Brock, and Parkes 2005) and appears unaffected in hypercapnia (Brown, Barnes, and Mündel 2014). Heart rate variability appears unaltered in both, although both low- and high-frequency components of HRV increase with added carbon dioxide compared to normal and reduced carbon dioxide levels (Pöyhönen et al. 2004). Electroencephalogram measures are altered in hypocapnia and hypercapnia, despite no difference in cognitive performance (Bloch-Salisbury, Lansing, and Shea 2000). Considering the above, measuring arterial PCO₂ is likely as important as oximetry to determine physiological and functional states.

3.4. Hypobaria

Cockpit barometric pressure can range from sea-level to ∼35,000 ft equivalent depending on the aircraft and flight profile. The primary concerns with hypobaria are pathologies of silent bubble formation, decompression sickness, and hypoxaemia above ∼20,000 ft (Webb and Pilmanis 2011). In contrast to arterial PO₂ and PCO₂, which can be detected by chemoreceptors in the body, there is no physiological baroreceptor of the external environment. Compared with normobaria, hypobaria is suggested to increase the ventilatory dead space (Ogawa et al. 2019; Savourey et al. 2003), which could increase the gradient between the end-tidal and arterial PCO₂ to increase ventilation and further reduce arterial PCO₂ in hypoxic conditions (Coppel et al. 2015). Hypobaria may also interact with carbon dioxide and hypoxia to blunt CBF and ventilation, which could impair cerebral oxygen delivery (Aebi, Bourdillon, Kunz, et al. 2020). Further, hypobaria appears to increase HR and marginally reduce HRV (Aebi, Bourdillon, Bron, et al. 2020); however, it does not appear to affect EEG power amplitudes (Kraaier, Van Huffelen, and Wieneke 1988). Repetitive changes in barometric pressure could also influence physiological responses; for example, venous gas emboli may increase or be more persistent when shifting between high and low altitudes (Ånell et al. 2020), although this is mitigated when breathing 100% oxygen (Ånell et al. 2021).

3.5. Gravitational and accelerative forces

An increase in gravitational forces, particularly in the +G_z axis (cranial-caudal direction), affects the cardiorespiratory system and haemodynamics. Loading in the +G_z direction beyond G-tolerance reduces blood flow to the eyes and brain (i.e. stagnant hypoxia) to cause tunnelled vision, grey-out, black-out, almost loss of consciousness (A-LOC), and gravity-induced loss of consciousness (G-LOC). The physiological response depends on the magnitude of total +G_z exposure (i.e. rate of onset, sustained duration, and peak magnitude) (Shender et al. 2003; Whinnery and Forster 2013). However, there is an initial functional buffer period of ∼5 s within which G-LOC does not occur (Whinnery and Forster 2013). A pulmonary perfusion-ventilation mismatch may form due to pulmonary atelectasis and shunting of deoxygenated blood in dependent portions of the lung. Blood pools in the lower limbs (i.e. below the heart) to reduce arterial blood pressure and cause central ischaemia. Ventilation and ventilatory heterogeneity increase (Borges et al. 2015), HR increases (Ueda et al. 2015), and HRV declines (Pipraiya, Tripathi, and Dogra 2005). Peripheral blood oxygen saturation (Eiken et al. 2017) and CBF decline (Eiken et al. 2017), reducing cerebral tissue oxygenation (Smith et al. 2013). Inflation of the AGS abdominal bladder can also contribute to hypoxaemia by facilitating compression atelectasis (Eiken, Bergsten, and Grönkvist 2011). Also, if the AGS deflates after the functional buffer period is depleted, the risk of G-LOC increases as cerebral tissue deoxygenation is accelerated (Eiken et al. 2017; Eiken and Grönkvist 2013).

3.6. Whole body vibration

Whole body vibration (WBV) produced by the aircraft ranges in frequency bandwidths to elicit various physiologic and symptomatic responses (Ballard, Madison, and Chancey 2020). Vertical vibration appears most influential (Ballard, Madison, and Chancey 2020) and can increase fatigue, drowsiness (Azizan et al. 2017), discomfort, and pain, particularly in the lower back (Bovenzi and Hulshof 1999). However, vibration duration, intensity, frequency, waveform, magnitude, and variation are also important factors (Conway, Szalma, and Hancock 2007; Kjellberg and Wikström 1985). Acute (∼2 min) WBV can enhance cognition (Regterschot et al. 2014), but can also cause impairment with increasing vibration duration and intensity (Conway, Szalma, and Hancock 2007). Whole body vibration increases ventilation (Hood et al. 1966) (and therefore the likelihood of hypocapnia), HR (Hood et al. 1966), and cerebral tissue oxygenation (Maikala, King, and Bhambhani 2005). Whereas, HRV declines (Zhang et al. 2018), although HRV can increase at low frequencies (Jiao et al. 2004). Some physiological perturbations return towards baseline with increasing duration (Kjellberg and Wikström 1985; Maikala, King, and Bhambhani 2006), suggesting vibration may become passively accommodated.

3.7. Heat and hypohydration

Heat stress is more likely than cold stress in military aviation due to the heat generated by the aircraft and the requirement to wear multiple layers of clothing and equipment. Heat stress causes discomfort, fatigue, and impairs mood (Simmons et al. 2008). An increase in core temperature can also impair cognition (Qian et al. 2014; Simmons et al. 2008). Heat stress may progress to cases of heat exhaustion and heat stroke if not managed appropriately (Epstein, Druyan, and Heled 2012). Increased sweating aims to dissipate heat through evaporative cooling; however, when sweat is trapped under layers of clothing and protective equipment, evaporative cooling is prevented and can cause discomfort. Sweat loss can lead to hypohydration, which may impair cognition when fluid losses exceed 1% of body weight (Lieberman 2007). Both heat and hypohydration also reduce G-tolerance (Nunneley and Stribley 1979). As the magnitude of hypohydration increases, the body’s ability to dissipate metabolic heat to the environment is reduced, thereby limiting heat tolerance (Sawka, Cheuvront, and Kenefick 2015). Heat stress increases ventilation (Qian et al. 2014; Worley et al. 2021) and HR (Bruce-Low, Cotterrell, and Jones 2006; Yamamoto et al. 2007); whereas, HRV (Bruce-Low, Cotterrell, and Jones 2006; Yamamoto et al. 2007) and CBF (Brothers et al. 2009; Nelson et al. 2011) are reduced. Hypohydration can also exacerbate the physiological effects of heat (Sawka, Cheuvront, and Kenefick 2015).

3.8. Sleep loss and circadian desynchronisation

Inadequate sleep (i.e. <7–9 h per night) and circadian desynchronisation is common within military aviation due to demanding operational requirements and travel across time zones. Sleep restriction is more common than total sleep deprivation, but can produce similar degradations in cognition and mood when sleep debt accumulates (Dongen et al. 2003). During sleep loss or circadian disruption, fatigue, sleepiness, and drowsiness increase, and cognition is impaired (Caldwell et al. 2019). Physiological perturbations to sleep loss accumulate across days and vary within the day (i.e. circadian effect), which can be subtle and modulate responses to other stresses. Sleep restriction does not appear to affect HR, but does decrease HRV (Dettoni et al. 2012); whereas, total sleep deprivation may decrease HR (Chen 1991). Cerebral blood flow tends to decline following sleep loss (Poudel, Innes, and Jones 2012; Zhou et al. 2019) and this effect may be negatively correlated with drowsiness (Poudel, Innes, and Jones 2012), although CBF has also been shown to increase following sleep deprivation (Elvsåshagen et al. 2019). Cerebral tissue oxygenation appears to decline during task performance following sleep restriction (Miyata et al. 2010; Yeung et al. 2018) and is positively correlated with sleepiness (Nishida et al. 2017; Suda et al. 2008).

4. Physiological episodes in military aviation

Physiological episodes did not initially feature in military aviation accident classification (Gibb and Olson 2008), but are now focal (Elliott and Schmitt 2019; National Commission on Military Aviation Safety 2020). A PE is any abnormal physiological condition, including illness and injury, consequent to the flight environment (United States Department of Defence Inspector General 2021). There are seven types of PEs, which relate to (1) oxygen availability and ventilation; (2) decompression illness; (3) barotrauma; (4) acceleration effects; (5) spatial disorientation and visual illusions; (6) toxic fumes, smoke or liquid exposures; and (7) other causes (United States Department of Defence Inspector General 2021). Inflight PEs manifest when a pilot’s physiological requirements are mismatched with the function and performance of the aircraft’s LSSs. Despite common elements between PEs, they do not always manifest in a consistent manner as differences occur between airframes, flight profiles, equipment malfunctions, and pilots. Therefore, without real-time physiological monitoring, reported PEs are difficult to corroborate and explain (i.e. Unexplained Physiological Episode [UPE]). Instead, they are retrospectively diagnosed by pilot recall, engineering investigations of aircraft LSSs and pilot equipment, and expert opinion.

From 2013 to 2018, 718 and 699 PEs were reported by the United States Air Force and the United States Navy and Marines, respectively. These incidences were predominantly in fighter (e.g. F-15, F-16, F-18, F-22, and F-35) and training (e.g. T-6, T-45, and T-38) aircraft with integrated wearable LSSs (National Commission on Military Aviation Safety 2020). Similar trends were previously reported by the Royal Australian Air Force, particularly hypoxia-like symptoms due to possible malfunctioning or misuse of oxygen equipment (Cable 2003). Although pilots operating large transport and reconnaissance aircraft are at lower risk of PEs, they still occur. Hypoxia-like events, such as those recently described in the Royal Air Force Typhoon aircraft (Connolly et al. 2021), are also potentially attributed to physiological perturbations other than hypoxia, such as hypocapnia. Factors, such as dehydration, spatial disorientation, motion sickness, or hypoglycaemia could also be implicated in PEs. Since physiological symptoms often present as mild, non-specific, and transient, it is possible that minor PEs remain undetected; therefore, PEs may be more common than reported and the risk of being physiologically compromised is higher than acknowledged. Due to the widespread involvement of numerous non-specific causes, PEs appear to be aircraft, LSS, and flight profile agnostic.

5. Integration of physiological monitoring systems

Physiological monitoring of military pilots will require a multi-modal suite of wearable and, probably, non-contact sensors. Sensors will need to be integrated within the aircraft and directly onto the pilot. Data from the sensors must be validated under the extremes of the dynamic flight environment as part of a layered system that will also monitor aircraft and LSS performance. Devices need to function during fluctuations in barometric pressure, G-forces, vibration, temperature, humidity, and in the micro-environment under the multiple layers worn by pilots (Phillips 2019). Whilst physiological responses can be determined to isolated stressors in laboratory settings (e.g. hypobaric chamber or human centrifuge), the full complement of environmental, physical, and psychological exposures during a single flight or multiple flights cannot be captured in these controlled environments. It seems the most relevant measurements are arterial blood gases, tissue oxygenation, haemodynamics, cardiorespiratory function, autonomic activity, and electrophysiological responses. By combining several physiological indices, a pilot’s cognitive and functional state can be more accurately determined.

The most promising wearable sensors for aviation applications include pulse oximetry, functional near-infrared spectroscopy (fNIRS), respiratory gas exchange, electrocardiogram (ECG), and EEG. Previous research has tested some sensors in the flight environment, such as fNIRS, to demonstrate reduced cerebral tissue oxygenation during aerial combat manoeuvres (Kobayashi, Tong, and Kikukawa 2002); however, these are typically prone to error. Alternatives, such as transcutaneous measurement of blood gases are also being considered (Shykoff et al. 2021), yet their validity in the laboratory (Lambert et al. 2018) and aerospace (Shykoff et al. 2021) environments appear poor. Based on a 2016 report by Defence Research and Development Canada, the United States military seems to be most proactive in developing physiological monitoring systems for pilots and are proceeding with operationalising advances in knowledge and technology (Burrell et al. 2016). The National Aeronautics and Space Administration have also recently published a series of studies evaluating ventilatory and respiratory changes using the VigilOX (Cobham Mission Systems, United Kingdom) sensing system during flight with varying profiles, equipment and regulator settings in high-performance aircraft (National Aeronautics and Space Administration 2020). Nonetheless, it is uncertain as to which sensors should comprise a multi-modal physiological monitoring suite to minimise intrusiveness, maximise comfort, remain non-interfering with emergency procedures, and generate the most informative data.

6. Non-contact physiological monitoring systems

Non-contact physiological monitoring systems could also be integrated within military aviation; however, there are inherent limitations on their value and they may not match the precision of wearable sensors. It is more likely that non-contact systems would supplement physiological data acquired from wearable sensors (Kutilek et al. 2019). As some wearable sensors may interfere with pilot comfort and performance, passive and unobtrusive systems, which can detect subtle changes in pilot physiology are highly desirable. Non-contact systems typically focus on the measurement of cardiopulmonary and respiratory responses, specifically the five vitals (i.e. HR, respiratory rate, body temperature, blood pressure, and S_pO₂), for which comprehensive reviews have been published (Leonhardt, Leicht, and Teichmann 2018). Video (e.g. head and body movement and facial expressions) and voice analysis, oculometry, ballistocardiography, and thermography, are also increasingly integrated into operational contexts. For example, subtle changes in pilot appearance and skin tone can detect changes in blood flow, gaze, and blink rate to determine workload and fatigue. The extent to which non-contact systems can be integrated will vary with aircraft design, mission demands, and pilot clothing and ensembles. The performance of non-contact systems to measure vitals may also vary between individuals; for example, pilots with variations in facial hair, skin tone, and head coverings may require individualised models as compatability may not be guaranteed in a generalised system. Non-contact systems may also require higher rates of data processing, which would increase the time of analysis, particularly when changes are subtle and need to be amplified to increase the signal to noise ratio.

7. Analysis and interpretation of data

Monitoring pilot physiology must provide an operational or training benefit, otherwise, it is superfluous to the flight environment. The goal of a monitoring system is to predict a pilot’s cognitive and functional state, and their capacity to perform optimally within the environment they are operating based on their physiology. Model outputs would therefore need to transform raw data into pragmatic outputs to guide pilot decision making and behaviour, which would be critical when a pilot’s cognitive and functional capacity is compromised. Physiological monitoring, in conjunction with aircraft and LSS monitoring, could also help identify equipment malfunction that may require corrective action. Physiological data would be used for diagnosis in incident and accident reports to corroborate pilot symptoms. This may also overcome barriers from pilots refraining to report PEs in concern of being medically grounded if there is a perception that PEs are their fault,and not the machine or training. Misleading or erroneous measurement of physiological data cannot be tolerated if used for predicting a pilot’s cognitive and functional state; therefore, measurement and analytical deficiencies need to be eliminated before the integration and implementation of physiological monitoring systems to promote operators’ trust.

Physiological sensors are sensitive and susceptible to different biological processes and physical conditions. This may increase the risk of predictive models being contraindicative to what the pilot is actually experiencing. The unpredictability and potentially extreme conditions of flight may also increase noisy data. Model algorithms would therefore be required to monitor data quality, extract important features, and identify and amplify subtle signals to construct a robust, accurate, and informative output. Predicting cognitive and functional state from a pilot’s physiological data is complex and unlikely to be possible from a single variable or sensor. Therefore, the fusion of data from multiple sensors is warranted, but this multiplies potential challenges as each sensor has inherent limitations, measurement performance (e.g. accuracy, reliability, resolution), and sampling frequencies. Aligning time-course changes between sensors can also be difficult, particularly if timestamps are inconsistently applied. Many of these challenges have recently been detailed in a study remotely monitoring physiological responses to extravehicular tasks during a simulated Mars mission (Hill and Caldwell 2022).

Behaviour results from the complex interplay of prior training and experience, as well as heuristics, and environmental and social cues (Johnson et al. 2020). Therefore, changes in physiological indices do not directly infer behaviour or decision-making. Behavioural responses to stressors, particularly in extreme environments when cognition is compromised, can also be illogical and augment, rather than ameliorate, the physiological stress, which would exacerbate cognitive and performance impairments (Costello, Wilkes, and Tipton 2021). The requirement to implement novel tasks may also increase a pilot’s cognitive load to cause errors if the tasks exceed their capabilities. Further, performance varies between and within individuals, which may affect the reliability of models, which was also demonstrated in the prior mentioned simulated Mars mission study, with the authors consequently emphasising the need for individualised and automated modelling tools (Hill and Caldwell 2022). Whereas, physiological responses to isolated stressors may be more easily predicted and conform to typical patterns, particularly if sufficient baseline data has been acquired. Considering the gap between physiology, cognitive and functional performance, and behaviour is yet to be bridged, generating inferences from physiological monitoring systems in operational environments would most likely be initially rudimentary.

8. Human–machine teaming

Newer generation aircraft will likely possess increased data handling and communicative capabilities as network-centric missions are of increasing priority. Integrating pilot physiological and (predicted) cognitive and functional status may eventually be another component for which the conceptual framework has been comprehensively discussed elsewhere (Lim et al. 2018; Liu et al. 2016; Pongsakornsathien et al. 2019). Advances in aircraft design will need to better accommodate and adjust to the changing state of the pilot if aircraft are to remain human-rated. This may include automating LSSs and stratifying critical tasks to enhance performance and safety when the pilot is compromised. Dynamic task allocation requires the aircraft’s systems to adapt and reconfigure to optimise the pilot’s performance within the confines of any impairment and ensure recovery. As pilots must remain engaged, task automation must also not lead to complacency and loss of situational awareness. A harmonious relationship is required between the pilot, aircraft, and environment, and rather than being top-down supervisory control, it is a collaboration through human-machine teaming. As aircraft become technologically more advanced and operational environments evolve, the criticality of understanding the responses and limitations of the human-machine interface becomes increasingly urgent.

9. Revising pilot physiological symptom recognition training

Military pilots are to undertake aviation medicine training at least once every 5 years according to the North Atlantic Treaty Organisation (NATO) Standardisation Agreement (STANAG) (NSO 2018) and Air Force Interoperability Council (AFIC 2022). Training includes exposures to stressors, such as hypoxia, changes in barometric pressure, and for jet pilots, +G_z forces. The purpose of hypoxia recognition training, using either or a combination of a hypobaric chamber and/or normobaric mask-on modalities, is to familiarise individuals with their constellation of personal hypoxia symptoms, experience the speed of onset and insidious nature of hypoxia, observe hypoxia-induced cognitive and psychomotor impairment in others, and practice using equipment and implementing emergency recovery procedures (Shaw, Cabre, and Gant 2021). If physiological monitoring systems were integrated into the aircraft, pilot training would require evolving to train how they fly to promote effectiveness and safety. Pilots would need to know how to interpret the outputs from models or how to respond if aircrafts are reconfigured. This also implies that pilots should be exposed to stressors in training that they are most likely to encounter during flight, rather than non-specific and misleading stressors that may increase flight safety risks. Presumably, this may demand the development of new training equipment to simulate real-world physiological events to avoid broadly applied training modalities. Biofeedback during training could also offer additional benefits, such as expediting skill acquisition and mastery (Ring et al. 2015). It is also plausible that pilots will ultimately have a better understanding of their physiological responses to aviation-related stressors and how these correlate to their performance under varying conditions, and are therefore able to to better prepare for missions and trainings.

10. Conclusions

Military pilots risk their lives to meet training and operational demands whilst flying aircraft that may stress them to their physiological limits. The physiological extremes of military aviation demand the integration of LSSs, yet even these systems can fail to meet demand. This uncoupling of advances in aerospace engineering and the understanding of how it affects the human piloting the aircraft clearly requires urgent addressing. Whilst the issue has been recently placed under the spotlight, the extent to which efforts are made to prevent PEs will assumedly be balanced with ensuring immediate capability is not compromised. Laboratory research also does not provide a true representation of the physical, psychological, and environmental stress pilots must endure in the real world. The data from these studies can only be considered preliminary and must be built upon in real-world settings or simulated environments. Moreover, some research assessing isolated or combinations of physiological extremes could have ethical challenges, particularly if requiring invasive procedures. In these cases, rodent or animal models could facilitate our initial understanding, as demonstrated in G-force research (Nishida et al. 2016).

Physiological episodes have grounded aircrew and aircraft fleets, cost billions, and have been unnecessarily fatal; however, eliminating the risk of PEs in military aviation is impossible. Understanding past disasters will also not prevent all in the future. Having an objective, real-time physiological data on the pilot could help to develop aircraft and LSSs. This may be acquired via a physiological monitoring suite, including wearable and non-contact sensors that do not interfere with the pilot or mission. Physiological monitoring would be part of a multi-layered system including additional data sources, such as aircraft performance, environmental conditions, and mission complexity. Due to the difficulty in predicting pilot behaviour and decision making from a physiological state, model outputs must be interpreted with caution to avoid unnecessarily compromising capability and safety. Sensors suitable for the flight environment that can be integrated within the pilot’s clothing and ensemble will need to be developed. An industry-military-academia interaction would help to accelerate advances in this technology and, once paired with advances in data fusion technologies, will enable Air Forces of Allied Nations with new capabilities.

Acknowledgements

The views expressed in this manuscript are those of the authors and do not reflect the official policy or position of the New Zealand Defence Force, Royal New Zealand Air Force, United States Government, United States Department of Defense, or the United States Department of the Air Force.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.