Localisation & Mapping with VizioSLAM

The advent of autonomous vehicles (AVs) has brought the need for precise localisation and mapping into sharp focus. With most AV implementations today, localisation and mapping are foundational to the AV’s ability to navigate and interact safely with its environment. But, is there a better approach? Let’s explore.

The Necessity of Accurate Localisation and Mapping

For an autonomous vehicle, understanding its precise location in the world and having a map of its surroundings is crucial, as without it, the vehicle would have no ability to navigate to a destination. However, in the context of most modern AVs, the level of mapping and localisation accuracy required is significantly greater than what you may expect from Google Maps on your smartphone.

Localisation

Localisation refers to the vehicle's ability to determine its position relative to a map. Unlike traditional GPS systems that offer accuracy within meters, AVs require centimetre-level precision to ensure they don’t stray outside of their lane or take the wrong turn in a complex driving environment.

Mapping

While a typical map shows routes, lanes and sometimes buildings, maps for AVs typically require much more detailed information, including road layouts, traffic signs, lane markers, and other critical infrastructure. Such maps are often referred to as High Definition or HD maps.

Current Technologies in AV Localisation and Mapping

Example Illustration of a HD Map

Real-Time Kinematic (RTK) and Differential GNSS (DGNSS)

Overview:

Standard GNSS (Global Navigation Satellite System) signals can have errors due to atmospheric conditions, satellite clock inaccuracies, and other factors, resulting in positional inaccuracies. To correct these inaccuracies, ground-based stations known as base stations are used. These stations are at known, fixed locations and continuously monitor GNSS signals to calculate the difference between its known precise location and the location indicated by the GNSS signals. This data is used to generate correction information, which is incorporated in RTK or DGNSS receivers to enhance localisation accuracy.

Challenges:

While RTK and DGNSS can provide centimetre-level localisation accuracy, they suffer from a variety of issues:

  1. Infrastructure Dependency: The effectiveness of such systems depends on the availability, proximity & maintenance of base station networks, which can be a constraint in terms of scalability and coverage.
  2. Latency in Correction Data: There can be a delay in receiving correction data, which can affect the accuracy in highly dynamic driving scenarios.
  3. Reliability in Diverse Environments: RTK & DGNSS systems often struggle in urban environments or enclosed spaces (such as multi-story or underground car parks) where buildings and other objects can obstruct or interfere with received signals. In such cases, localisation accuracy is often significantly reduced.
  4. Cost: RTK & DGNSS systems require specialised equipment and are considerably more expensive than standard GPS/GNSS units.

HD Maps

Overview:

High Definition (HD) maps contain comprehensive, highly accurate data about roadways and their surroundings, including:

  1. Lane Information: Exact lane positions, widths, and the direction of travel.
  2. Traffic Signs and Signals: Locations and details of traffic signs, signals, and road markings.
  3. 3D Environmental Data: Information about buildings, trees, curbs, and other static objects around the road.
  4. Road Geometry: Precise details about the road's curvature, slopes, and elevation changes.

In AVs, HD Maps are used to provide centimetre-level localisation accuracy, even in the absence of a GPS signal. This is done by matching data from onboard sensors, such as LiDAR and camera, with the information stored in the HD map to determine the vehicle’s precise location.

Challenges:

While HD Maps overcome the issue of providing centimetre-level localisation accuracy in environments without a GPS signal, they also suffer from a variety of issues:

  1. Maintaining Up-to-Date Maps: The dynamic nature of road environments means that HD maps need to be regularly updated to reflect current conditions, such as changes due to construction or road works.
  2. Data Processing and Storage: The detailed information contained in HD maps requires significant data processing capabilities and storage, both on the vehicle and in the cloud.
  3. Integration with Sensor Data: Effectively integrating and cross-referencing HD map data with real-time sensor data is a complex task that requires sophisticated algorithms and computing power.
  4. Scalability, Coverage & Cost: Creating and maintaining HD maps for extensive road networks is a massive undertaking, requiring substantial resources and coordination.

Introducing VizioSLAM

VizioSLAM (Simultaneous Localisation & Mapping) enables centimetre-level localisation and HD map generation, with minimal hardware complexity and cost. Let’s see how it works:

  1. Leveraging the high spatial resolution of Provizio’s 5D Perception sensors, novel software techniques are used to extract visual reference points from radar point cloud data.
  2. As the vehicle moves within its environment, this data is extracted from each new point cloud frame.
  3. With every new frame produced, specialised algorithms compare the point clouds frame by frame and use the visual reference points as a means to calculate changes in the velocity and rotation of the vehicle relative to its surrounding environment.
  4. With our high-speed, on-the-edge perception process, VizioPrime and VizioPlex sensors can feed these algorithms with enough data of sufficiently high resolution to provide centimetre-level localisation accuracy, without the need for HD maps or GPS.

The Provizio Advantage

  1. Reliability in Diverse Environments: VizioSLAM is not negatively affected by complex environments such as dense urban areas or enclosed spaces. As a result, it is a more robust solution than RTK or DGNSS.
  2. Scalability: VizioSLAM works on-vehicle and in real-time, without the need for external correction signals or HD maps. As a result, the system can be applied in any environment, without any prerequisites for HD Maps or base station networks.
  3. Cost: Without the need for costly DGNSS sensors or to maintain up-to-date HD maps, VizioSLAM is a once-off cost for OEMs.
  4. Integration: With our unique on-the-edge perception process, VizioSLAM can be easily integrated as part of a larger AV stack, with minimal overhead in terms of compute or network demands.
  5. Mapping: Like LiDAR, VizioPrime & Plex sensors can be used to generate high fidelity environmental maps. As a result, OEMs can use our sensors to build their own crowdsourced mapping solutions, and reduce reliance on third-party HD map providers.

Conclusion

Accurate localisation and mapping are vital components in the development and operation of autonomous vehicles. While current technologies like DGNSS, IMU & HD maps have provided an initial starting point for the development of AVs, the challenges in scaling these technologies limit their appeal for mass market applications.

With VizioSLAM, we believe we have developed a compelling alternative for both accurate localisation and mapping, which can cater for the needs of mass market adoption. If you’re interested in learning more about VizioSLAM, visit the VizioSLAM webpage, where you can find more detailed technical information and demonstration videos. Alternatively, if you’d like to get in touch with us directly, please visit Contact Us.

Our Approach to Sustainable Development

At Provizio, our commitment to innovation doesn't just stop with our technology. It extends into the very way we work. To address the challenges associated with an ever increasing need to boost workplace productivity, while simultaneously giving employees a fair work/life balance, we needed to take a novel approach to product development that wasn’t bound to traditional rules of project management. So, let’s dive in!

Taking a Step Back

As a startup organisation, it’s easy to fall into the trap of following “the industry standard” when establishing the foundations of a workflow management process. With so many things to build and so little hands to build them, startups often find themselves so entrenched in the engineering process that they fail to see the bigger picture of why.

At Provizio, we understood that in order to supplant industry leading competitors, we needed to focus as much on innovation within our development process as in our products. In this regard, an “if it works for them it can work for us” approach was not going to provide the momentum we needed to ultimately build a product that could entice customers away from the grip of trusted suppliers. Hence, a key approach to continued innovation within Provizio, is the ability for us to regularly step out of the trenches and take a candid look at whether our process and priorities are truly aligned and in service of our overall mission.

The Intersection of Workplace Psychology and Productivity:

In today's rapidly evolving business landscape, where factors like remote work, reduced hour work weeks, flexible working hours and artificial intelligence are hot topics among employees, many companies face the challenge either of holding their ground with tried and true processes, or taking a chance to embrace the unknown in the hopes of coming out on top.

At Provizio, we believe our workforce is the heart of our success and as such, we spend a considerable amount of time in pioneering innovative work practices to boost productivity, creativity, and employee satisfaction to ensure that our heart continues to beat strong and true. These practices, deeply rooted in the principles of workplace psychology, are reshaping how we think about work, collaboration, and productivity.

Agile Development

Illustration of people brainstorming, planning and thinking of ideas on a Kanban board

Agile principles are the bedrock of our project management approach, allowing us to respond swiftly and effectively to the ever-evolving demands of our industry. However, while many organisations get bogged down in agile process documentation, arguing between the pros and cons of the now countless agile frameworks available today, at Provizio we take a different approach.

Our work practices are guided mainly by values outlined in the original Agile Manifesto, whose deliberately vague guidelines left the door open for organisations to customise a workflow approach that worked best for their unique structure, products and workforce. In this respect, we made our own interpretations of the values of the Agile Manifesto and built a framework tailored for our specific needs.

When it comes to agile frameworks, Provizio takes a little inspiration from almost every one.

These, and many other frameworks like LeSS, Nexus, The Spotify Model and more, all fed into the development of our current agile process which, like our products, improves and evolves incrementally based on regular feedback.

Overcoming Resistance

A frequent barrier to agile development, often voiced by those exposed to more traditional or waterfall based development styles, is the alleged lack of evidence to support the view that agile will work in their specific industry. Now, while there have been many studies to support that agile values such as individual autonomy and incremental development in the workplace have been linked to higher job satisfaction, improved performance, increased creativity & enhanced product quality, there will always be a sense of skepticism when adopting a new process.

However, the key to overcoming such resistance is to ensure that employees have a sense of choice and initiative. In agile, nothing is fixed and scrutiny is welcome to provide the insights needed to improve and adapt the process to suit every unique case. In giving people an outlet to express their feedback over time, with the knowledge that their feedback is heard and actioned, this often leads to higher levels of intrinsic motivation, which is a key driver in producing higher quality work.

Conclusion

At Provizio, our approach to work mirrors our approach to technology - innovative, agile, and always focused on the human element. By embracing agile principles and granting autonomy to our employees, we're not just building advanced technologies; we're shaping a work culture that is as progressive and cutting-edge as our products.

Together, these practices form the cornerstone of our success, driving us towards our ultimate goal – making roads safer for everyone.

Enabling 5D Perception with MIMSO & SPTDMA

The advancements in radar technology, particularly the development of 4D radar, have revolutionised various fields, including automotive safety, aviation, and weather monitoring. At the core of these sophisticated systems lies significant innovations in digital signal processing (DSP) and antenna design, both of which have been a central focus of Provizio R&D for the past number of years. In this post, we’ll explore the current challenges faced by 4D radar systems, and how Provizio technology pushes the bounds of what’s possible in the space.

What is 4D Radar?

4D radar is an evolution from traditional radar systems, adding the dimension of elevation to the existing coordinates of range, azimuth, and velocity. This additional data dimension allows for more precise tracking and identification of objects in three-dimensional space and their speed.

Key Components of 4D Radar Systems

  1. Transmitter and Receiver Arrays: Modern 4D radar systems typically use phased array antennas, which consist of numerous small antenna elements. Each element can be controlled independently, allowing for electronic beam steering without moving the antenna mechanically.
  2. Frequency Modulated Continuous Wave (FMCW) Technology: Many 4D radar systems utilise FMCW technology, where the radar signal’s frequency is varied over time. This allows for the simultaneous measurement of range and velocity.
  3. Multiple Input Multiple Output (MIMO) Techniques: MIMO technology is used to improve the resolution and accuracy of a radar system. By transmitting multiple signals and receiving their echoes through multiple channels, MIMO enhances target detection capabilities and enables improved spatial and angular resolution.
  4. Fast Fourier Transform (FFT): FFTs are employed to convert time-domain radar signals into the frequency domain. This transformation is essential for determining the range and velocity of targets.
  5. Doppler Processing: By analysing the Doppler shift in the frequency of returned signals, the speed and direction of an object can be determined.
  6. Angle of Arrival (AoA) Estimation: Sophisticated DSP algorithms estimate the AoA of radar signals, which provides information on the object’s azimuth (horizontal location) and elevation.

Challenges with Existing 4D Radar Technology

  1. Computational Complexity: The high volume of data and the complexity of the algorithms require significant processing power.
  2. Real-Time Processing Requirements: The requirement for real-time signal processing for applications like autonomous driving and collision avoidance is technically demanding. There is a constant need for faster and more efficient processing algorithms.
  3. Antenna Design: The complexity of designing phased array antennas for 4D radar systems is significant. These antennas must accurately manage beam steering in three dimensions, which can be technically challenging and expensive.
  4. Size and Integration: Incorporating 4D radar systems into smaller platforms, such as those required for mass market automotive or mixed mobility applications, can be constrained by the size and power requirements of the hardware.
  5. Manufacturing Costs: The cost of producing sophisticated phased array antennas and other components of 4D radar systems can be high. This impacts the scalability and adoption rate in various industries.
  6. Complex Algorithm Development: Developing algorithms that can effectively interpret the complex data from 4D radar systems is challenging. These algorithms must be robust, accurate, and able to differentiate between a wide range of target types.

Taking 4D to the Next Dimension

Provizio 5D Perception® Elevation Point Cloud Showing a Pedestrian Overpass

Key to 5D Perception® is our advancements in radar resolution and range. Digital radars in vehicles today generally have a single radar chip, with 3 transmit channels and 4 receive channels that allow them to deliver a virtual antenna array of 12 elements. In radar, the number of elements is like the pixel count for cameras and larger virtual apertures are preferred to improve resolution. Today’s radar has a typical detection range from 100m to 150m. These sensors enable the basic and dependable adaptive cruise control (ACC) and Automatic Emergency Breaking (AEB) on vehicles. However, to enable safe autonomy Provizio brought learning from advanced aerospace applications and developed a dedicated cost-effective solution for the automotive vehicle market.

Previous to Provizio our team built products that NASA described as “awesome” and the world’s leading autonomous driving group described as “the gold standard” in automotive Radar. That experience led us to envision a completely different, patent protected (7 at the time of writing), approach to software defined imaging radar.

MIMSO®

MIMSO® (Multiple Input Multiple Sparse Output) is a software defined active antenna technique, which embeds proprietary surface-mount technology (SMT) integrated circuits (ICs) into a novel planar antenna design. This allows us to lower the receive path noise floor and discriminate more of the radar beam by essentially recycling parts of the radar beams that have traditionally been filtered. On the transmit path, our SMT ICs allow us to carry out instantaneous beam switching, which multiplies our transmit channels and further increases our resolution. Ultimately, MIMSO allows us to extract more than 30x the resolution from each physical radar channel.

SPTDMA™

Our proprietary software modulation technique, SPTDMA™, is a new form of multiplexing that enables unparalleled 6K resolution out to over 600m in all weather conditions, with the added benefit of protecting against interference from other sensors.

Industry Limitations

In radar systems, there are different modulation schemes that can be used to transmit and receive signals from multiple targets. One such scheme is Time Division Multiple Access (TDMA), where each transmitter transmits a signal on separate time slots. This allows the received signals from each transmitter to be easily separated with minimal computational overhead. However, using TDMA modulation comes at a cost of reduced maximum unambiguous velocity measurement, since the frequency at which measurements can be performed is limited by the time delay introduced by the allocation of time slots for each transmitter.

To address this issue, another modulation scheme called Doppler Division Multiple Access (DDMA) can be used. In this scheme, each transmitter transmits simultaneously (eliminating the time delay) and a series of algorithms are used to recover the velocity of a detected object by correctly matching each received signal to the transmitter that sent it. However, this approach is limited by challenges in accurately determining the velocity of targets and resolution limitations for objects that are closely spaced. Moreover, the addition of complex algorithms to match transmitter/receiver signals requires sophisticated signal processing that can lead to increased computational demand and pose challenges for real-time processing.

The Benefits of SPTDMA

Provizio’s patented SPTDMA (Slow Phase Time Division Multiple Access) solution combines TDMA and DDMA techniques by splitting multiple transmitters into smaller sub-arrays, with each sub-array transmitting simultaneously on an allocated time slot. In this way, while there is a limitation on maximum unambiguous velocity compared to a pure DDMA system, we can increase the number transmit channels to improve angular accuracy and range, while keeping computational complexity low enough to enable real-time, on-the-edge processing.

The Competition

The incumbent Tier 1 radar manufacturers are all developing 4D radar to enable Level 3 driving. They have very little differentiation and are almost all building forward facing 4D radars with 192 virtual antenna elements, delivered using 4x COTS radar chips. A big improvement, but nowhere near AV requirements.

There are also other competitors working to develop super-resolution radar. These solutions generally fall into one of two camps:

The Provizio Advantage

Conclusion

Provizio has invested significant resources in pushing the bounds of what’s possible with 4D radar hardware and software technology. With our patented MIMSO® and SPTDMA™ systems, Provizio offers finer resolution, faster processing speeds, and more accurate target detection and tracking, delivering compelling solutions to enable the transition to L3+ autonomy and the future of automotive safety.

Mastering Work-Life Balance at Provizio

Provizio was established not only to do great work, but also to be a great place to work. A key goal of the company is to be an industry leader in driving work-life balance and positive mental health initiatives to support all employees, both inside and outside of work. Provizio also understands that to attract and retain the best talent, providing and supporting a flexible working environment that can adapt with the needs of employees is of vital importance.

Our Initiatives

The 4-Day Work Week:

Provizio is a firm believer in “it’s not about how many hours you work, it’s about what you do with the hours that you do work”. While the logic of this sentiment is something many people agree on, the concept of a 4-day work week is one which still scares many employers. Despite various research studies highlighting the potential benefits of a 4-day week, the thought of losing eight hours of precious development time each week proves too difficult a pill to swallow for many organisations. However, at Provizio, like with anything we do in our organisation, we approached the 4-day week like an experiment - create a hypothesis, test the hypothesis and evaluate the results.

In adopting a 4-day work week, a critical enabling component was to first understand how our existing 5-day process would adapt to ensure productivity wasn’t compromised. This included evaluating meeting schedules, understanding common productivity disruptors, assessing our team alignment processes, and ensuring that individuals were empowered to enhance their individual day to day productivity.

After a successful trial, it was amazing to see just how much our workflow process changed in accommodating a reduced hour workweek. New initiatives like “Deep Dive Days”, meeting blocks and asynchronous chat and video updates enabled vast reductions in daily disruptors, while an extra day of rest each week enabled employees to spend more time with family, friends, or on their favourite hobbies, leading to increased productivity and job satisfaction.

Remote Work and Flexible Hours:

Remote work and flexible working hours are other critical aspects of Provizio’s work-life balance approach, aligning with current workplace psychology research. One study found that employees working remotely reported higher job satisfaction and reduced attrition rates, owing to employees being able to create a personal work environment conducive to their productivity. Flexible working hours also play a significant role, as evidenced by a 2016 study which found that flexible schedules led to reduced stress and burnout.

In our own experience, remote work and flexible working hours are a critical enabler of enhanced creativity and productivity for most employees. Our internal studies showed that many employees use both flexible working hours and remote work to their advantage, taking time during the traditional workday for running errands or minding children, and using quiet time later in the evening for deep focus.

In addition, many people have different circadian rhythms, with some preferring to do their most complex work in the morning and others preferring to work at night. With a flexible and remote work approach, employees are free to utilise the work day to their advantage, maximising productivity while minimising distraction. This approach isn’t just good for our employees; it’s beneficial for our business. A rested and fulfilled team brings more energy and creativity to their work, driving innovation and helping us stay ahead of competition.

Evidence of a Triple-Peak Workday - Source: Microsoft

Productivity Boosters:

In addition to the initiatives outlined above, we found the below techniques to be very helpful in improving overall productivity. However, it is key to remember that not every approach is suited to every person or role, so please take the time to assess what may work best for your unique situation.

Provizio Productivity Boosters

Conclusion

Provizio’s adoption of these progressive work practices is not just a response to modern trends but a well-considered strategy informed by extensive research in workplace psychology. By fostering autonomy, embracing agile methodologies, and promoting a healthy work-life balance through innovative scheduling, Provizio is not only enhancing employee satisfaction and retention but is also driving the creation of higher quality products.

In the evolving landscape of work, Provizio stands as a testament to the power of aligning workplace practices with targeted, informed research, showcasing how a modern approach to work can lead to substantial benefits for both employees and the organisation.

It’s Time to Start Treating Road Deaths Like a Global Pandemic

At the time of writing most of the world is in lockdown. Our more vulnerable citizens are cocooning in their homes and our communities and politics are divided. It’s a global pandemic, caused by an uncompromising killer.

According to the World Health Organization (WHO), there have been 1.2 million Covid-19 deaths worldwide, 44 million cases and the cost to the global economy is estimated to hit $1 trillion.

The world has rallied. United in the belief that this level of carnage is not acceptable we are taking draconian measures to stop the spread and investing billions in the science that will ultimately save millions of lives.

The automotive pandemic

According to the WHO, each year there are over 1.35 million road deaths worldwide, over 50 million people maimed and the cost to the global economy is estimated to be over $2 trillion (Link to the study here: WHO)

The numbers quoted above have not changed for more than 20 years. Unless action is taken, they will continue their current trajectory. Preventable human error is responsible for more than 90% of road accidents.

International regulatory policy for the last 20 years has focused on trying to make us better, more responsible drivers. This approach has failed consistently.

The focus by automakers on driver comfort and reactive safety measures has also, predictably failed.

Autonomous driving will ultimately solve the problem but has failed to move beyond level 2 autonomy despite $80 billion invested and will continue to do so while the roads are shared.

The vaccine

We started Provizio to solve the global road death pandemic. We believe that with the right focus, robotics and drivers can work together to reduce both road deaths and accidents to zero. 1.35 million to zero drives everything we do.

We are building augmented, guardian angel technology that will make us all better and safer drivers. We are using unparalleled ‘beyond line-of-sight’ sensor technology coupled with artificial intelligence ‘on-the-edge’ to perceive, predict and prevent accidents.

We are partnering with the top automotive OEMs and Tier 1’s to bring this capability to the masses.