Archives
Contribute
| Technology - Quantifying Performance Using Human Function Signatures: Part II |
Shawn McGuan
11/25/2006
With human engaged products, it might be a good idea to consider the human as early in the design process as possible. This would lead to deeper design insight, fewer prototype iterations and subsequently a better product delivered to the market faster. How could this be done?
For an example of a product engaged inside a human, consider a total knee replacement system (see BRG Newsletter Volume 3 ). Manufacturers of this device survive on how well they can optimize the function of their products while maintaining good product robustness and durability.
By generating a detailed leg, complete with ligaments, bones and muscles, the CAD model of the implant concept may be imported and exercised. Here, the signature could be kinematic motion such as how the components move while the model climbs stairs. With this methodology, the baseline could be a healthy knee as well as existing and competitive designs.
Figure 4. Musculoskeletal human leg model with a total knee replacement system. This model is used to develop the motion and force signature of the implant system. Many simulations may be performed for popular activities such as running, squatting, lunging, etc., providing an ideal benchmark environment to optimize function. The curve displays one aspect of the performance signature - the tibial internal rotation with respect to flexion angle for the simulation and an intact (healthy) knee as a reference.
We have seen vehicle designers gain greatly from this process. Consider a motorcycle (see BRG Newsletter Volume 4) controlled by a virtual human rider. The simulation would involve a virtual motorcycle vehicle concept model, before any parts were created or assembled. The human rider would "steer" the model through a specific course graphing out a signature of human torques necessary to stabilize and control the motorcycle. This signature can be developed into a reading of the comfort and controllability of the proposed vehicle.
Figure 5 Human model "steers" a full vehicle motorcycle model through a slalom coarse. Ride signature is reported in terms of control effort by the human model which is linked to the vehicle model using a closed loop control system. The curves display one aspect of the signature indicating the amount of effort required to manage the vehicle through the virtual test ride.
The human-centered design cycle
The examples listed above include the human response to the product design in a form of a definable signature which can be used to rank designs, competitors, etc. This quantitative feedback becomes the missing function element in the design stage as displayed in figure 6.
Figure 6 The modern product development process from figure 1 with the functional component of the virtual design/testing cycle augmented with human simulation.
In these examples, as is in many other, a clear definable methodology for this human functional component can be ascertained. These steps include:
Develop the appropriate model - Start simply and increase complexity of the model to produce accurate, consistent results without being overly complex.
Make sure the model is valid - Obviously the model must produce accurate results in order to accurately predict human response. There are a variety of approaches to assure validity not only for the model but for the range of events the model will simulate. It is our experience that in the product development process, rather than developing a model with exquisite detail for a specific individual, it is better to develop an "averaged" model which reflects a specific population that accurately reproduces trends.
Define a small set of test events - Develop a small set of key product usage events to cover the range of the product.
Define the performance signature - From the simulations, extract a concise data measures to characterize performance.
|
You may also access this article through our web-site http://www.lokvani.com/
