Sensor Technologies for Automotive Systems

Existing LiDAR systems are being used for innovative uses from security to mapping to industrial automation. One sector with specifically high levels of interest and development is the mobility market. LiDAR scanners are crucial components in prototype systems for autonomous vehicles, and also in current systems for traffic sign recognition, adaptive cruise control (ACC), blind spot detection, collision avoidance systems and lane departure warning.

Image Credit: Shutterstock/posteriori

None of the above LiDAR-based systems can operate in the absence of a key component, which is their sensors, or the “eyes” of the system.

The focus of this article will be on how design engineers for original equipment manufacturers (OEMs) of LiDAR systems can choose from various sensor technologies.

Background

Mobility light detection and ranging (LiDAR) systems must sense the environment quickly and reliably, assembling as detailed a picture as possible of immediate surroundings and the road ahead. Far in advance: systems fitted in fast-driving cars should be able to “see” a minimum of 150 m (almost 500 feet) ahead, as well as detect objects as small as 10 cm (around 4 inches) in height.

Figure 1. LIDAR-scanner in action. Image Credit: First Sensor

All the above points present challenging technical requirements for the sensors of the systems.

The task at hand demands complementary yet independent sensor systems, with assured environmental qualification and functional safety. For instance, to accommodate both heat from other system components and environmental heating, units have to be rated for operating temperatures from −40 to 125 °C (−40 to 257 °F). To be able to “see” the signal through any distracting backdrop, sensors should possess an ideal signal-to-noise ratio. Moreover, as optical detectors have to be prepared to handle differing levels of environmental light, sensors must possess a broad dynamic range.

(It should be noted that the term “detector” might refer only to the photoelectric detecting element, whereas the term “sensor” refers to the detector as well as associated electronics that offer functions such as connectivity. At times, the terms are used interchangeably.)

Apart from basic physics, designers of LiDAR systems should also take into account basic economics. All components in the car have to be predominantly cost-effective. For practical purposes, the best cost/performance ratio is the key to the best technology.

All the existing automotive mobility systems that use long-range LiDAR are “scanning” devices that move the laser beam gradually over the entire scene. Effective range using current technology: 30 to 300 m (about 100 to 1000 ft). Nearly all the systems are built using 905 nm lasers, which can be obtained at low cost in high volumes, emit invisible beams, and use high power for short pulses (for instance, 75 W peak for 5 ns) — offering an ideal power-to-cost ratio. These lasers have been widely applied using advanced, low-cost silicon detector technology.

Selecting the Best Sensor Technology

As the industry advances, design engineers are using various different sensor technologies for LiDAR mobility systems. Each has its own benefits and disadvantages, as shown below.

Silicon PIN Diode Detectors

These silicon-based detectors have a structure formed of three types of semiconductors layered together — P-type/Intrinsic/N-type.

They demonstrate optimum dynamic range, with the potential to deal with largely varying amounts of light. For instance, they have the ability to detect the reflection of a distant object, even upon being subjected to direct sunlight. Moreover, they are comparatively economical.

Yet, they do not have the ability to offer the high levels of signal-to-noise performance and bandwidth required by majority of the sophisticated mobility LiDAR systems. Lastly, they are neither very fast nor very sensitive.

Silicon Photomultiplier (SiPM) and Single-Photon Avalanche Diode (SPAD) Detectors

The makers of these solid-state, silicon-based sensors originally created these for specialized, small medical and scientific applications. In the recent times, they have made efforts to try them out in the larger LiDAR market.

Although the functioning of these sensors is analogous to that of APDs (discussed below), they are optimized for exceptionally high internal amplification or gain, enabling them to detect even trivial amounts of light. They also function very fast. Lastly, they are compatible with commonly available CMOS technology, and hence can be coupled with associated electronics on the same chip.

Yet, when compared to the sensitivity of APDs, the sensitivity of the single-photon counters of these sensors is considerably lower. Therefore, they must rely on extremely high multiplication. Regrettably, during the multiplication process, noise is added, which often considerably degrades the signal-to-noise ratio. Their amplification mechanism is also subject to false triggers resulting from high temperatures.

Presumably, the most critical disadvantage of these sensors is that their high gain comes at the expense of saturation problems.

Primarily, the sensors should handle laser light reflected from objects ahead. Moreover, several LiDAR systems specify scanners that have wide fields of view. This places an enormous amount of added light on an SPAD or SIPM sensor. In addition, certain phenomena regularly experienced in LiDAR mobility environments — such as high-beam headlights, bright sunlight, or other LiDAR systems — can saturate the sensor with higher light levels than it can deal with, even when optical filters are used.

With continuous development efforts to offset their disadvantages, these sensors are often used for different LiDAR applications. However, until now, their saturation problems and other challenges described above have prevented them from being the detectors of choice for scanning long-range LiDAR.

Figure 2. Comparison of detection technologies. Image Credit: First Sensor

Indium Gallium Arsenide (InGaAs) Photodiode Detectors

These sensors are often applied at small sizes in telecommunications glass fibers, but have been only recently applied for LiDAR, apart from specialized aerospace or military applications. This technology rejects traditional silicon-based construction for InGaAs material.

Since laser systems are specifically designed for its higher spectrum (1550 nm, versus 905 nm for the other sensors discussed in this article), this technology should be highly sensitive, and possess the ability to discharge more power. Thus, it has the ability to enable development of an automotive LiDAR system with a longer range compared to the majority of other sensors.

However, the performance of an InGaAs detector can be considerably degraded by even negligibly higher-than-normal ambient temperatures. The sensor may require an external cooling system, even under temperate climates.

Furthermore, its base material is considerably more costly compared to largely used silicon substrates. Moreover, the fabrication of InGaAs sensors in large sizes for LiDAR use would involve far more complex fabrication than silicon designs. Until now, it has not been possible to successfully produce in high commercial volumes.

Lastly, as this technology is new to the field of automotive LiDAR, OEMs will have to be ready to spend considerable effort, time, and cost when attempting to build an innovative LiDAR system around any InGaAs detector.

Avalanche Photodiode (APD) Detectors

These silicon-based photodetectors were primarily perfected for industrial and military applications. They operate by allowing incoming photons to initiate a charge avalanche, multiplying gain by their internal amplification mechanism. With their absorption-optimized structure, they convert nearly 80% of the 905 nm reflected light of a laser into photoelectric current. The outcome is greatly increased sensitivity.

Apart from their prominent sensitivity, APDs have minimal saturation, an ideal signal-to-noise ratio, and high speed. They are also one of the most low-cost sensor technologies in the market.

A potential disadvantage of APDs is that they use specialized bipolar technology that is not compatible with ordinary CMOS fabrication. Therefore, it is possible to source them only from a handful of suppliers. Moreover, they cannot be coupled on the same chip with their associated CMOS electronics.

However, packages with sensor and electronics can be fabricated on chips located close to each other by experienced suppliers. Both can be enhanced for top-rated performance, without any compromise. For instance, specially designed transimpedance amplifiers (TIAs) that have tailor-made gains and bandwidths can be used to complement for an APD sensor array — to transform the photocurrent to voltage, and to condition the signal that enters into the system for increased gain. This can improve the performance, specifically in low-light conditions.

APDs are manufactured using standard, high-productivity, commercial production processes and their ability has already been demonstrated in a broad range of systems on the road.

Fundamentally, when performed in the proper way, they integrate demonstrated performance with an attractive cost. APDs are, at present, the detectors of choice for automotive long-range LiDAR and are crucial components in several existing most sophisticated mobility systems.

Figure 3. APDs. Image Credit: First Sensor

Selecting the Best Sensor Supplier

Upon determining the apt sensor technology, LiDAR system designers are still left with the difficulty of choosing the apt sensor supplier.

It is necessary to assess the candidates with caution. Do they possess the potential, technology, and proficiency to adapt their systems and sensors to the individual requirements and markets of an OEM? Will they work hand in glove with the OEM team on the design, production and scheduling to guarantee a winning time to market?

Insist on Experience

In case a sensor vendor will have to spend time in speeding up its development, production, automotive qualification, and other processes, the LiDAR mobility OEM will fail in the race for the fastest time to market.

Experience is gained by sensor suppliers by performing the work. A good candidate will have already used their detector/sensor technology for mobility applications. This may include tailor-made and standard APD design; tailor-made and standard packages, dies, and modules engineering and manufacture; and top-rated electronics.

A perfect supplier will have a confirmed track record, with products such as automotive-grade APDs and related electronics already being used by leading LiDAR OEMs.

Assess for Integrated Manufacturing

It is necessary for designers to prioritize a supplier with appropriate technical benefits, such as highest sensitivity and lowest noise. However, it is also necessary for them to search for a sensor maker maintaining a holistic control of its domain.

It is important to implement the complete production process as a coherent whole — from chip processing to prefabrication of sensor systems. If a supplier manufactures all important components in-house, it can guarantee the long availability of all OEM products for aftermarkets and series production.

Check Customization Capabilities

In order to emerge as a successful LiDAR system maker, it is vital to realize the best cost/performance ratio, which can assist in differentiating a specified system from the remaining overcrowded market. Off-the-shelf sensors may not be suitable. Rather, components must usually be tailor-made to be exactly suitable for a chosen system design.

System makers have to find a responsive and agile sensor supplier. In several instances, a supplier has to work with OEM designers to tailor-make the sensor and related electronics for the stringent possible integration with the remaining system — and, hence, for best-in-class performance.

Examples: the team has to establish sensor geometries that are suitable for a specified choice of lenses, optimize dimensions, and to otherwise adapt to the configurations of each distinctive optical design. The team has to determine the ideal channel count — the number of signals received in parallel — to increase the scanner’s spatial resolution. Moreover, it should tailor-make packaging for the shortest possible interfaces between electronics and sensor.

Lastly, an outstanding supplier must offer sensors that have sophisticated technical advantages such as multi-pixel homogeneity. Under real-world use, non-homogeneous and/or diversely sourced photodiodes react differently to ambient temperatures, thereby considerably degrading the performance of the LiDAR scanner. On the other hand, multi-pixel homogeneity can offer the most stringent signal information distribution, even at peak distances.

Ask About Automotive Qualification

A superior sensor supplier should already be aware of the “rules of the road.” They should be experienced with up-to-date robustness validation, automotive qualification, and characterization regulations and standards.

Examples include ISO/TS 16949 automotive-certified production and testing, as well as AEC-Q 102 and 104 automotive-qualified APD array packaging. The supplier should be in a position to apply these and other appropriate standards to all its components and manufacturing facilities, to adhere to regulations and to assist system OEMs in avoiding liability.

Increasing regulation is unavoidable. Suppliers should exhibit proven adherence through documented best practices, like the stringent self-qualification pioneered by Ford Motor Company in its Q Program.

Look for Future-Proof Support

It is also necessary for a supplier to exhibit a proven track record of quality as well as delivery performance, and also strong support levels, from initial development to maintenance service.

Sensor design should be taken into account from the start of system design. The earlier the sensor supplier is made involved by an OEM, the easier and faster the entire design/manufacturing process will be — and the better will be the performance of the resulting LiDAR system.

Lastly, a supplier must always be looking forward to future advancements in this rapidly changing field. The right sensor manufacturer will have an innovation roadmap of anticipated regulatory, business, and technological advancements to come, to assist system makers to navigate through this rapidly changing market.

Conclusion

The eye of every LiDAR system is its sensor. It is possible for system designers to choose from various competing sensor technologies. Various design engineers believe that APD sensors offer the ideal combination of cost and performance. Makers of LiDAR systems should also take into account various factors in choosing their sensor supplier — such as customization capabilities, experience and automotive qualification expertise.

With the continuous advancement in LiDAR and other mobility technologies, making the right sensor choices evidently paves the way forward.

This information has been sourced, reviewed and adapted from materials provided by First Sensor Inc.

For more information on this source, please visit First Sensor Inc.