Wireless sensor platform eases quest for parking

3Cruising around looking for a parking space wastes gas, increases pollution, and frustrates drivers short on time. Using a wireless sensor network to accurately determine parking vacancies enables cities to
decrease congestion and manage their resources more efficiently.

Throughout the world, quality of life is depreciated by atmospheric pollution and congested roads, which result in lost time for drivers and wasted fuel. The European Commission estimates that economic losses due to traffic delays total €150 billion per year in Europe. Across the Atlantic, things are not much better. In 2000, the Texas Transportation Institute reported that the 75 largest metropolitan areas in the United States suffered 3.6 billion hours of vehicle delay, causing the loss of 21.6 billion liters of fuel and $67.5 billion in productivity.

The need to search for available parking spaces – often involving lengthy periods of driving around idly and racing other vehicles to open spots – is a significant contributor to widespread congestion and a major cause of stress for motorists. Based on calculations by UCLA professor of urban planning Donald Shoup, vehicles searching for parking in a small business district in Los Angeles cruise an estimated 950,000 miles a year, wasting about 95,000 hours (11 years) of drivers’ time and 47,000 gallons of gasoline, while producing 730 tons of CO2 emissions. And in Barcelona, Spain, a million drivers spend an average of 20 minutes every day looking for a parking spot, producing 2,400 tons of CO2.

Technologies for parking management

To reduce traffic congestion produced by drivers searching for free parking spots, cities are turning to technologies that can provide accurate information on where to find available parking spaces, thereby minimizing delays, helping traffic flow more smoothly, and even enabling drivers to book parking spaces using applications on their smartphones or via in-vehicle infotainment systems.

This type of solution works like an indoor parking structure system, which uses ultrasonic sensors, GPS receivers, and cellular networks to find empty parking spaces and then relays this information to drivers using Internet maps, navigation systems, and displays installed in the garage. These classic systems are wired, requiring sensor installation on top of the vehicles, and must be positioned in range of the data receivers.

When moving to a city scenario, this approach is not possible. Wiring the parking lots in a city incurs significant installation expenses, and the types of sensors used indoors can produce false positives caused by other vehicles or objects near the lot under control. In addition, parking sensors must be robust enough to be buried under parking spaces and handle wide fluctuations in environmental conditions.

As opposed to the infrared or ultrasound sensor systems used in garages, parking in the city requires mesh networks to reduce the number of receivers. Deploying mesh networks for parking management presents challenges, such as addressing the potential of false rejection due to not high enough variation in the magnetic field above the mote with a vehicle parked in the spot. Furthermore, variations in temperature, presence of external fields, or small changes in the magnetic field of the Earth can cause a small oscillation in the reference value from which the state of the lot is determined. This can affect the detection threshold and thus increase the probabilities of false detection or false rejection.

Deploying a sensor network

As an alternative to conventional wired systems used in indoor parking structures, Libelium developed smart parking sensor technology (shown in Figure 1) as part of the company’s Smart Cities platform. The concept calls for deployment of a with sensors placed inside the concrete, one per parking spot. These battery-powered sensors detect the arrival and departure of vehicles within the parking lot and transmit the information wirelessly to different panels displaying the number of places available.

21
Figure 1: The smart parking wireless sensor network detects the arrival and departure of vehicles in a parking lot.
(Click graphic to zoom by 1.9x)

The hardware used in this solution is Waspmote, a modular platform for building . The platform comprises:

  • A Waspmote board with microcontroller, memory, battery, accelerometer, and sockets for add-on modules
  • Open-source API and compiler
  • Array of wireless communication modules offering a choice of protocol versions, radio frequency, and range up to 12 km
  • Wireless modules supporting Bluetooth, GPRS, and GPS
  • Variety of sensor boards enabling the measurement of gases, physical events, agricultural, and smart metering parameters

Measuring 73.5 mm x 51 mm x 1.3 mm and operating within a -20 °C to +65 °C temperature range, the sensor board was designed specifically for this smart parking application. The overall platform addresses a number of technical challenges unique to wireless sensor networks deployed in a city parking setting, including detection, , outdoor protection, power consumption, and maintenance costs.

Detection

One of the main functions of a smart parking system is detecting the vehicles occupying spots. Indoor systems typically use ultrasound or infrared sensors that detect presence and movement close to them, which would generate false positives in a busy street. To increase detection accuracy, the Libelium wireless sensor network uses an electromagnetic field sensor or “mote” buried under each parking space. The sensor detects variation in the magnetic field above caused by the chassis of a vehicle parked in the lot.

Each sensor comprises a thin film of permalloy whose resistance is a function of the magnetic field through it. This film is integrated in a Wheatstone bridge of resistance between 1.2 kW and 2.2 kW, thus rendering a voltage of 20 mV with a supply voltage of 5 V between the sensor’s two output terminals. Each Waspmote board contains two of these sensors as well as four small magnets to help minimize the effect on the measurement of small variations in the magnetic field, such as those of the Earth’s magnetic field. An instrumentation amplifier amplifies and filters the sensor’s output to prevent glitches caused by external magnetic fields, and an -to-digital converter reads the outputs directly in the analog input pins ANALOG1 and ANALOG7.

Network

To avoid problems in reaching the main receiver, all the sensors communicate with ZigBee/802.15.4 radios in a wireless mesh network. The ZigBee/802.15.4 protocols work in the same frequency band as (2.4 GHz); thus communication is free and does not require a license purchase. Without this, it would be impossible to completely monitor large areas in which there might be hundreds or even thousands of devices.

Using a mesh topology allows the network to be more scalable and robust since new nodes are automatically recognized and faulty ones do not affect the rest of the network. Additionally, 900 MHz and 868 MHz radios are compliant with this solution for covering longer distances between motes. For 2.4 GHz ZigBee/802.15.4 connections, mesh networks are implemented with routing motes located in street lights. For the lower-frequency radios, parking sensors communicate directly with the gateway since the propagation distance is longer.

Outdoor protection

Protecting sensors in parking scenarios is critical, as the motes must last for a long time buried in asphalt, where they are subjected to high pressure from the parked vehicles and face exposure to water from puddles on the pavement. To address these concerns, the sensor nodes are enclosed in a PVC casing rated at IK10 for mechanical impact protection and IP67 for ingress protection. The use of PVC ensures that radio communication is not hindered. Figure 2 shows the sensor inside the enclosure installed in a parking lot.

22
Figure 2: Sensors installed in a parking lot must be able to withstand high pressure from parked vehicles as well as exposure to harsh weather.
(Click graphic to zoom by 1.9x)

Power consumption

When the board is disconnected and the components are inactive, power consumption is null. When the board is on, consumption varies between 7.2 mA maximum and 6 mA minimum, depending on the state of the sensors. The electric resistance the sensors show, and thus the current that flows through them, is a function of the magnetic field through the permalloy film that forms the sensor. Power consumption thus depends on the state of the parking spot and is not controlled by the user.

The board can be programmed to switch off any of its components if needed to save energy. For example, the ZigBee radio can be turned off and activated only when a change in the state of the parking lot must be transmitted.

Maintenance costs

Because the devices are ultra-low-power and programmable Over The Air (OTA), installed sensors do not need to be accessed for many years. Motes only need to transmit when a parking event – a vehicle arriving or leaving a space – takes place. With suitable batteries, a sensor can operate for five years before it needs to be physically accessed for battery replacement. OTA programming enables the software for entire networks to be upgraded efficiently over the radio network without digging up parking spaces. These two features reduce the maintenance involved in smart parking sensor networks, enabling hundreds of nodes to be readily deployed.

Smart parking, smart city

This smart parking wireless sensor network was introduced earlier this year at Sensors Expo & Conference, and its first deployment is in SmartSantander, a unique city-scale experimental research facility that supports smart city applications. Located in the Cantabria region on the north coast of Spain, the facility comprises more than 20,000 sensors and fosters the development and deployment of “Internet of Things” technologies in an urban setting. The Network Planning and Mobile Communications Group from the University of Cantabria helped test and fine-tune the parking sensor for this application.

The first phase of the project consists of 600 parking sensors using 802.15.4 radios with the Digimesh protocol. Temperature, luminosity, and noise sensors are included in the routing motes in the street lights to provide additional information to the city government. By supplying accurate information on available parking spaces, this wireless sensor network offers a comprehensive parking management solution that cuts pollution and congestion for a smarter, healthier, more connected city.

Alicia Asín is cofounder and CEO of Libelium, a wireless sensor networks hardware manufacturer. She previously worked in computing security with open-source tools and is currently a member of the Wireless Sensor Networks and Mesh Networks international research groups. She has a Computer Engineering degree from the Polytechnic Center of Zaragoza, Spain.

Libelium +34 976 54 74 92 info@libelium.com www.libelium.com