Truly intelligent precision positioning has arrived, and it’s able to perform autonomous and simultaneous optimization, which will affect a range of global industrial sectors.
SCOTT JORDAN, PHYSIK INSTRUMENTE
Over the span of 40 years, the field of precision positioning has vaulted forward in deeply beneficial ways. Submicron-resolution linear encoders, for example, once available only in the costliest stages, are now affordably integrated into mechanisms as large as gantries and as small as matchbook-sized nanopositioners (Figure 1). New motor concepts abound, ranging from the familiar DC servo, stepper, and linear motors to novel technologies such as high-speed, ultrastable resonant piezo motors and diminutive stick-slip stages with nanoscale positionability over many millimeters of travel.
Figure 1. Gantry positioning systems today are available with linear motors and absolute encoders, providing nanometer-scale resolution (a). Matchbox-sized nanopositioning stages based on piezo motor technology (b). Courtesy of Physik Instrumente.
Meanwhile, applications as complex as silicon photonics packaging automation, micro-optical assembly tooling, and nanorobotics all benefit from new combinability of axes and mechanisms of various types. Controls technologies have broadened and specialized but have also adopted extensible interfacing architectures, including both proprietary buses and open approaches such as EtherCAT (Ethernet for control automation technology), a high-speed, real-time industrial networking and distributed-control architecture based on ubiquitous Ethernet hardware. These interfacing architectures provide highly distributed capability where each axis and functionality is optimized for its task, resulting in an assembly with no compromises.
A study in obedience
As advanced as today’s precision positioning mechanisms and controls have become, they still basically do what they are told. A position is commanded and achieved. A sequence of motions is specified, interleaved with metrology and other processes, and it proceeds according to script. In a strictly operational sense, these systems differ little from the positioning systems of 1985. Advancements, however, have come in resolution, speed, compactness, and similar physical attributes, and in controller attributes such as scriptability, data recording, and other features.
The emerging field of intelligent positioning systems refers to industrial microrobots that automatically seek the optimum position of optical elements, sensors, and other components. Courtesy of Physik Instrumente.
Take a high-altitude view and another trend begins to manifest — one with profound significance for applications. The broad tectonics of the positioning field, upon review, make it clear that positioning is suddenly becoming more intelligent as well. Specialized systems already have become available that not only achieve commanded positions and motion patterns with alacrity, but can determine, on their own, the optimal position and orientation needed for their application.
Trends in positioning
The intelligence trend started in the field of photonics process automation. Even at the dawn of the fiber optic field in the late 1980s there was a need to automatically align elements for test and packaging,accommodating device-to-device variations and fixturing uncertainties that could not be machined out or detected with even the best machine vision. This optimization was achieved by software.
In those early days, computers were used to sequence precise but dumb positioners along with optical-power metrology in ways that could eventually determine and achieve the optimal mutual orientation of photonic elements such as the then-novel single-mode fibers and laser diodes. The optimization sequence could be as simple as a brute-force raster scan of an area of interest to determine the position of a peak; or starting in the early 1990s, a digital gradient search could be performed to climb the coupling profile efficiently and even provide tracking of thermal drift and curing stresses.
But the intelligence remained in the computer. The algorithms that performed the optimization resided in the software, and the throughput of the system was crimped by the ability of every positioning axis and metrology subsystem to respond to the computer’s moment-by-moment commands. At the same time, the sophistication of the software’s real-time capabilities was limited by the slowness of all the elements being orchestrated. Until recently, this remained the state of the art.
Exquisite alignment
Over time, more and more sophisticated optimization algorithms were developed, for example, to allow simultaneous optimization of multiple photonic elements — which included silicon photonic chips, fiber arrays, lenses, and diffraction gratings — across multiple channels and degrees of freedom. These algorithms use micron- or even nanometer-scale exploratory motions to measure the local gradient of a figure of merit such as optical power, and then the systems automatically follow that gradient to the point at which it falls to zero: the peak, the optimum, the desired position and orientation of the elements. This is a definitive and unique condition familiar to students of the Euler-Lagrange equation and — as it turns out — a condition universal to optimization challenges (Figure 2).
Figure 2. The unique parallel gradient search algorithm finds the signal maximum typically in less than 1 s, allows tracking and compensation for drift, and enables simultaneous optimization of multiple inputs, outputs, and degrees of freedom. Courtesy of Physik Instrumente.
In a groundbreaking advancement, these newly intelligent algorithms have recently deployed in the firmware of specialized controllers whose compute power handily exceeds that of the Cray supercomputers of those early days of photonics. Combined with multiaxis positioners of exquisite precision, these parallel, intelligent optimization technologies provide microrobotic functionality that reduces the time required to assemble photonic devices versus traditional, sequential-alignment techniques, and not by a small amount: Process time reductions of 99 percent or more are common. This has proven to be an economic enabler for the industry, and the systems have enjoyed considerable commercial success and many awards (Figure 3).
Figure 3. A photonics-enabled wafer prober integrates a multichannel photonics alignment system based on hexapods and parallel-kinematic piezo scanners for high-throughput optical probing of on-wafer silicon photonic devices. Courtesy of Cascade Microtech, division of FormFactor Inc.
Broadly reducing costs
But remember that high-altitude view. It turns out that this new class of intelligent positioning is suitable for many applications outside of silicon photonics. After all, many industrial deployments of ultraprecision motion control are aimed at peaking up some quantity or another as devices are assembled. The manufacturing of lasers is a good example. The output power of the laser must be optimized. Camera manufacturing is another example. In that field, image quality is the metric to be maximized. And so on. The devices being mutually aligned can even integrate their own sensors and logic, opening possibilities for coordinated or even collaborative robotic optimization on a submicron scale.
Examples abound. All involve a figure of merit that reaches a peak at the desired position and orientation. And similar bounties of process-cost reduction are seen when conventional micropositioning is supplanted by these new intelligent optimization engines combined with high-quality industrial multiaxis positioning mechanisms.
Furthermore, the pioneering parallelism that benefited silicon photonics production economics so dramatically — the ability to simultaneously align multiple photonic elements across multiple channels and through multiple degrees of freedom in one step — also applies to the general case of industrial precision assembly.
Intelligence via parallelism
The key is how the parallel optimization can replace time-consuming loops. A gimbaling optimization of a lens — for example, in θX and θY — formerly needed to periodically halt so the transverse alignment in XY could be corrected, and then the θX/θY optimization could start over again, and on and on in a lengthy loop until a global consensus optimum is achieved. Now, both optimizations can proceed simultaneously, yielding substantially faster global optimization and greatly reducing process costs.
Figure 4. Simultaneous alignment of multiple lens elements in multiple degrees of freedom in one step enables significant cost and time savings in the production process. Courtesy of Physik Instrumente.
This is broadly applicable to many manufacturing fields. All that is needed is for the optimization controller to directly receive an adequately fast figure of merit, and the optimization can then begin across all the involved channels and degrees of freedom. In familiar silicon photonics applications, the quantity being optimized is optical throughput, meaning power. So the signal conveyed
to the controller is typically the output of a high-bandwidth optical power meter or transimpedance amplifier. That is all the controller needs to intelligently determine the optimal position and
orientations of the elements it is positioning (Figure 4). Signals can be shared across controllers, again facilitating collaborative robotics in these applications. Digital format and calculated figures of merit are also easily accommodated.
Return to the new application examples cited above, and it’s easy to see how this capability maps to these new applications. In the case of a laser cavity, the mutual orientation of reflectors, gratings, and other constituents similarly must be optimized, and the figure of merit is optical output or some measure of modal quality or spectral purity. Dependencies between elements and geometrical dependencies for each element can be unwrapped automatically through the parallelized algorithm. In the case of a multielement camera, such as the billions of increasingly sophisticated smartphone cameras assembled each year, the metric of image quality can be a straightforward and fast calculation of image sharpness, such as a 2D FFT (fast Fourier transform) or modulation transfer function calculation. When conveyed to the controller at a high rate, this can similarly drive the simultaneous optimization of elements across multiple degrees of freedom, reducing or eliminating the need for those time-consuming process loops.
Optimization is universal
The pioneering parallelism that benefited silicon photonics production economics so dramatically — the ability to simultaneously align multiple photonic elements across multiple channels and through multiple degrees of freedom in one step — also applies to the general case of industrial precision assembly.
Key to all this is the fact that most figures of merit are substantially unimodal peaking functions near optimum, meaning they exhibit a hill-shaped profile that rises and then falls as the orientation of each element is exercised in each of its degrees of freedom. The same internalized mathematical algorithms that achieved the radical capability of simultaneous optimization in silicon photonics production and test apply in these additional applications as well. At the root of this capability is the novel, parallelized digital gradient search, a specialized category of gradient ascent algorithm that itself is a relative of the Euler-Lagrange equation. This is a highly generalized technology, hence its broad applicability. Importantly, a model of the coupling and dependencies need not be known, nor must the application be predictable or tightly reproducible.
A precipitous drop in cost
What this means for production economics across these new fields is still unfolding, but the promise is similar to what we have seen in silicon photonics process automation. Fields as diverse as lidar, lasers, the life sciences, data storage, quantum computing — virtually any field in which production economics are affected by lengthy process loops in pursuit of a global consensus optimization — will benefit.
Return once again to the high-altitude view, where a foundational shift in positioning technology has occurred — from its advancing history of positioners ever more precisely doing what they’re told, to a new paradigm of truly intelligent positioning. The era dawns where positioners can now perform autonomous optimization of process quantities, and the consequences for industry will be broad and deep.
Meet the author
Scott Jordan is head of photonics for Physik Instrumente (PI), and a PI fellow. A physicist, with an MBA in finance/new ventures, Jordan has made multiple contributions to positioning and optimization technologies; email: [email protected].