Continuous Product Improvements
This article is more on the theme of how much R&D and investment it takes to stay in the GNSS game. I was pleased to recently talk with Hemisphere, headquartered in Calgary, about some of the features of its new V102 system, with NovAtel in Calgary on its OEM product line evolution, and with Septentrio in Belgium about its family of dual-frequency receivers.
Hemisphere (when it was still CSI) fielded the two-piece Vector heading sensor in 2002. The initial objective was to provide heading back-up for ships’ gyros. The two small antennas were mounted in a single enclosure spaced about a meter apart, and the processing receivers were housed in a separate unit. The external antenna assembly was mounted on a mast or tractor roof, open to the elements, then lots of RF cabling to connect things together. Lots of RF cable means lots of connectors into which water or rain could find a way to accumulate and corrode important connections. Lots of cables added cost, and installation time was also expensive.
Several companies have struggled with GPS antennas on ships and the effects of multipath, which reaches these antennas off really good reflectors such as deck and superstructure.
So then Hemisphere went into product update round one, it found room to move the receivers into the external housing along with the antennas, and the system no longer had susceptible external RF cables or connectors. Performance improved and installation costs were reduced – good for the user, good for sales. With this V-100 product, Hemisphere was successful in penetrating the commercial and professional marine markets.
As sales of the V-100 moved along, Hemisphere was able to afford to get into product update round two. These updates further improved front-end performance, allowing tracking satellites at low-elevation angles and under foliage. Interference rejection was also improved to allow better integration with radios, cell-phones, and other transmitters.
Talking with NovAtel, I learned that the company has used an evolutionary development cycle for a number of years. Its first OEM receiver was fielded around 1994, a 10-channel L1 receiver known as “OEM1.” Each OEM1 receiver had two digital ASIC “engines” at the heart, each capable of tracking five signals. Total power consumption for the ASICs to track 10 channels was more than 2 Watts (200 milliwatts/channel). Over the next 16 years a new receiver emerged, on average, every two years with the current OEM6 120-channel multiple frequency GPS/GLONASS/Galileo receiver evolving through many generations of development. For comparison, the MINOS6 ASIC at the core of this receiver has 120 channels on a single ASIC, and total power consumption for that ASIC is 0.8 Watts (6.7 milliwatts/channel). Each phase of receiver development has become increasingly more complex as more and more integration has reduced size and power while increasing functionality. The complexity and size of the software core of these receivers quickly outstripped the capability of the ’90s processor, and today’s machine uses an advanced ARM derivative.
Each step along the development path has been driven by lessons learned in the marketplace and through customer applications. Interfaces have been added or modified to enable customer installations to become feasible or to work better; features have been introduced, expanded and refined; and new capabilities have been added. As the customer base has increased and many diverse applications have developed and been proven, the overall design has become increasingly more robust. Of course, as more functionality is added and complexity increases, the effort involved in ASIC, digital, RF, and software development has expanded and the test and qualification cycle has become more extensive. The growing NovAtel receiver portfolio and customer base has allowed increased sales to provide the internal funding needed to maintain the pace of this development cycle.
And finally, Septentrio provided me with a story about how its 2003 PolaRx2@ single-board receiver with multi-antenna capability evolved. The large number of channels and the flexibility of the ASIC (GReCo2) at the core of the original PolaRx2 platform, led to the Septentrio concept of adding multiple antenna inputs. Because of the inherent flexibility of the PolaRx2 architecture, an initial multi-antenna product was readily derived from the main single-antenna receiver platform, allowing for fast market introduction.
PolaRx2@ was a unique combination of a dual-frequency positioning engine (including RTK) with a single-frequency-based attitude calculation. The receiver was designed with one dual-frequency antenna input and two single-frequency antenna inputs.
Exposure with customers over the next 2-3 years provided many important practical inputs. One set of feedback allowed Septentrio to improve the single-frequency attitude algorithm, to make it more robust and to perform well in a variety of user environments.
A more crucial and fundamental feedback was that learned first-hand about the requirements and opportunities in using multi-antenna receivers for machine control. It became clear that a dual-frequency, dual-antenna receiver would have tremendous value for a slew of applications in earth moving and related machine-control applications.
Septentrio redesigned the existing PolaRx2@ receiver to drop one antenna input, and reconfigure the freed-up channels for dual-frequency operation on the second antenna as well. Combined with the algorithmic improvements already made on PolaRx2@ for fast acquisition and fast and stable GNSS-based attitude determination, this resulted in the unique 2006 dual-frequency RTK PolaRx2eH receiver design.
With the replenishment of the GLONASS constellation, and the use of GNSS receivers in ever more demanding environments, the need to add GLONASS became obvious. A new receiver was conceived based on a new ASIC (GReCo3), and the AsteRx2eH was introduced in 2009. More signals, higher update rates and lower latency, improved tracking with RTK performance — these have all made this latest version of Septentrio’s multi-antenna receiver even more popular than the previous products.
So these examples seem to indicate that technology and product evolution drive an ever-changing GNSS industry. These three leading manufacturers all appear to invest continuously to field ever evolving, improving products, learning from continuous customer feedback. As new applications are encountered, these manufacturers have responded by funding internal product improvements and have benefited from increased market penetration with improved products. The cycle never seems to end, and as long as there is continued market opportunity, the industry appears to be able to maintain the cycle of Continuous Product Improvement.
Tony Murfin
GNSS Aerospace
Follow Us