Originally published in 2014; rebuilt in 2026. This post originally contained only a video. The video featured Hugh Herr’s work on advanced bionic limbs. The article has now been expanded into a public education resource about how designers and engineers turn human needs into working systems.
Quick Read
Designers and engineers change the world when they translate a real human need into a solution that works under physical, biological, economic and ethical constraints.
One-sentence answer: strong engineering begins with the receiver, identifies the real problem, builds a model, tests it against reality, learns from failure and keeps improving until the solution is useful, safe and reliable enough for real human use.
The original video: Hugh Herr and bionics
The video linked in the 2014 post is Hugh Herr’s TED talk on bionic prostheses. Herr, a biomechatronics researcher and double amputee, demonstrates how engineering can combine mechanics, electronics, sensing, control and human biomechanics to build artificial limbs that support walking, running, climbing and dancing.
The educational value is larger than the device itself. Bionics provides a clear example of an engineering problem in which the machine cannot be designed separately from the human being who will use it.
Design starts with a human need
An engineer can build something technically impressive that solves the wrong problem. That is why useful design begins by defining the receiver and the desired outcome.
- Who will use the system?
- What does the person need to do?
- What currently prevents that outcome?
- Which constraints cannot be ignored?
- What would success look like to the user?
For a prosthetic limb, success is not merely “the motor moved”. The person needs comfort, control, balance, durability, manageable weight, safe loading, reliable attachment and movement that works in everyday situations.
Engineering is constraint management
Real engineering rarely asks for one variable to be maximised without cost. A lighter structure may be weaker. A more powerful actuator may consume more energy. A more complex sensor system may become harder to maintain. Higher performance may increase cost.
The engineering task is therefore to find a workable region inside many constraints rather than chase one perfect number.
| Constraint | Question |
|---|---|
| Physics | Can the structure carry the required forces? |
| Energy | How long can the system operate? |
| Human factors | Is it comfortable and understandable to use? |
| Safety | What happens when something fails? |
| Cost | Can the intended user reasonably access it? |
| Maintenance | Can it be serviced over time? |
| Ethics | Does the design preserve dignity, agency and choice? |
The design loop
- Observe: understand the real environment and user.
- Define: state the problem precisely.
- Model: predict how a possible solution should behave.
- Prototype: build a testable version.
- Measure: compare actual performance with the prediction.
- Fail safely: identify what breaks without harming the user.
- Revise: change the design based on evidence.
- Validate: test again under realistic conditions.
The loop is powerful because it keeps the design answerable to the world. If reality disagrees with the model, the model has to change.
Bionics is a systems problem
A modern powered prosthesis can contain structural components, joints, sensors, processors, control algorithms, batteries and actuators. The user also contributes motion, intention, balance and adaptation.
No single component explains the final outcome. The useful behaviour emerges from the whole system working together.
This gives students a general engineering lesson: when a product appears simple at the interface, complexity may have been moved behind the interface through good design.
Sensing: the machine needs information
A robotic or bionic system cannot respond intelligently without information about its current state. Sensors may measure angle, force, acceleration, contact, pressure or other variables.
The basic control problem can be expressed simply:
- What state is the system in now?
- What state should it be in?
- What action should reduce the difference?
- Did the action actually work?
This receive-predict-act-check loop appears across engineering, from thermostats and drones to industrial robots and medical devices.
Control: movement must remain stable
A machine can have enough power and still be unusable if its control is unstable. In a prosthetic limb, movement must respond quickly enough to the user while avoiding oscillation, sudden unsafe motion or delayed correction.
Control engineering therefore turns raw hardware into coordinated behaviour. The goal is not maximum motion; it is appropriate motion at the right time.
The human-machine interface
Medical and assistive engineering requires an especially careful interface because the user is not outside the system. The device touches the body, changes movement and may become part of daily identity.
A strong design therefore asks more than “Can this device function?” It asks:
- Can the person understand and control it?
- Does the device fit safely?
- Can it be worn for meaningful periods?
- Does it reduce or create fatigue?
- Can the user trust what it will do next?
- Can the person override or stop it?
Failure is part of engineering
Students sometimes believe good engineers avoid failure. In reality, engineering depends on discovering failure early enough that it becomes information rather than harm.
A prototype should reveal weak assumptions. A stress test should identify where a component breaks. A simulator should expose control problems before a human depends on the system.
The aim is not to eliminate all failure from experimentation. It is to contain failure inside safe testing environments until the system is robust enough for wider use.
Safety has to be designed in
Safety cannot be added as a final label after the exciting work is complete. Engineers need to anticipate failure modes from the beginning.
- What if a sensor gives the wrong value?
- What if power is lost?
- What if the network connection disappears?
- What if the user moves unexpectedly?
- What if two subsystems disagree?
- Can the system fail into a safer state?
In medical and assistive technology, the cost of failure reaches a human body. That makes safety architecture a core part of the design.
Accessibility is part of usefulness
A device can work beautifully in a laboratory and still fail as a social solution if very few people can access it. Cost, clinical support, repair, training, geography and healthcare systems all affect whether technology reaches the person who needs it.
This is a useful distinction between invention and impact. Invention proves that something can exist. Impact requires the surrounding route to work too.
Repair versus enhancement
Bionic systems also raise an ethical boundary question. A technology may begin as a way to restore lost function and later develop capabilities that exceed ordinary human performance.
That does not make enhancement automatically good or bad. It does mean society needs to ask who receives the technology, who pays, what risks are acceptable and whether new capabilities create unfair or coercive expectations.
Engineering and AI
AI can improve perception, prediction, optimisation and control, but it does not remove engineering responsibility. A learning system still needs defined operating boundaries, reliable sensors, appropriate fallback behaviour and a clear human authority structure.
A model can recommend an action. The larger system must still determine whether that action is safe, authorised and suitable for the actual user.
What students can learn from bionics
| Subject | Learning route |
|---|---|
| Mathematics | Forces, geometry, optimisation, signals and modelling |
| Physics | Motion, torque, energy and stability |
| Biology | Muscles, joints, gait and nervous-system interaction |
| Computing | Sensing, control, embedded systems and algorithms |
| Design | User needs, prototyping and iteration |
| Ethics | Access, agency, enhancement and safety |
| English | Explaining complex systems clearly to non-specialists |
The deeper design principle
The best engineering is not impressive because it contains many parts. It is impressive because the parts disappear into a useful human outcome.
That is why Hugh Herr’s bionics work remains a strong educational example. It connects a human need with biomechanics, electronics, materials, computation, control and design—then returns the result to the human being who must live with it.
Source
The original video is Hugh Herr’s TED talk, The new bionics that let us run, climb and dance, published in 2014.
Historical note: this page began in 2014 as a single-video post. The 2026 revision keeps that original object and adds a durable engineering-learning layer around human needs, constraints, sensing, control, safety, iteration and responsible innovation.