Enabling the next generation of AI powered exoskeletons

The International Federation of Robotics reports more than 542,000 industrial robots installed in 2024 and global deployments now exceeding 4.6 million units. At the same time, Health and Safety Executive data shows that manual handling accounts for around 17 per cent of all non-fatal workplace injuries in Great Britain.

Exoskeletons sit between these pressures, offering targeted physical support in areas where automation cannot fully replace human movement.

From assisted movement to intelligent support

Early exoskeletons prioritised lifting support, often at the expense of flexibility. Movement could feel rigid, which limited long-term use. That is now changing as AI enables systems to respond in actual time.

Traditional systems relied on pre-programmed assistance curves. These worked well in controlled tasks, but struggled with variation. AI-driven systems instead adjust output in real time based on movement, load and environment.

This shift is already visible in practice. In the UK, the ABLE Exoskeleton has been introduced into rehabilitation programmes, where it is used to support patients with spinal cord injuries and gait disorders through adjustable, personalised assistance. Its design allows clinicians to adapt support to each patient’s needs, reflecting a broader move towards systems that respond dynamically rather than follow fixed movement patterns.

Why AI changes motor requirements

AI-driven exoskeletons place new demands on motor performance. Instead of steady output, motors must handle rapid torque changes and continuous micro-adjustments. Load conditions can shift from one movement to the next.

Motors must also allow natural movement when assistance is low. High resistance can interfere with balance and increase fatigue. Low-friction, accruable designs help ensure the user remains in control.

Brushless DC motors are widely used due to their reliability and consistent output. Their high torque density allows engineers to meet performance targets without increasing system bulk. In practice, inconsistent torque delivery can affect gait stability, particularly under changing loads.

Closing the gap between thought and motion

Even small delays can disrupt balance and coordination. In highly precision-dependent applications such as surgical robotics, latency below 50 milliseconds is generally perceived as responsive, while delays above 100 milliseconds begin to affect control.

Although this benchmark originates from surgical systems, similar latency thresholds are relevant in other human-interactive robotics. In exoskeletons, such delays can reduce gait stability and make movement feel less natural. 

To mitigate these effects, most robotic systems rely on closed-loop control using multiple sensors. These typically include accelerometers, force sensors and position encoders, which provide continuous feedback to guide motor response in real time.

Some designs go further by predicting movement patterns. This reduces lag and improves flow between actions. 

Managing variable loads and thermal behaviour

Unlike fixed systems, AI-driven exoskeletons operate under highly variable duty cycles. Movement patterns change throughout the day, creating fluctuating loads and unpredictable demand on the motor.

Inconsistent torque delivery can affect gait stability, particularly when transitioning between movements. Reliable motor control is essential to maintain smooth and predictable motion.

Thermal behaviour becomes more complex under these conditions. Heat does not build evenly, increasing the risk of local hotspots. Efficient motor design and heat dissipation are critical for maintaining performance during extended use.

A modular future, powered by precision

Exoskeleton development is moving towards more adaptable systems. Future designs will adjust support based on user behaviour over time and this will improve both comfort and performance.

Modular design is also becoming more common. Components can be upgraded without replacing the full system as this extends product life and allows new technologies to be adopted more easily.

AI-driven exoskeletons are changing how people work and recover from injury. Their success depends on consistent, controlled movement under real conditions, placing motor performance at the centre of system reliability. 

To learn how EMS supports adaptive motion in exoskeleton design, visit the website and speak to the team about your application.

Dave Walsha is sales and marketing director at Electro Mechanical Systems (EMS)

www.ems-limited.co.uk