NIRVANA MERATNIA: In this lecture, I will talk about sensing and identification technology and its application in the logistics and transport sector. Why do we need sensing and identification in the first place? Because consumer and businesses would like to know about the quality and they would like to control the quality of the goods. There’s a need for anti-theft, for instance. There’s a need for on-time delivery. Faults and errors occurred during load, transport, storage, and delivery should be identified and reduced. The same applies to storage in compatibilities. They should be identified timely. So what are the mechanisms or what are the technologies used for sensor identification in logistic sector? Barcode is a commonly used method. It needs visual contact.
It can be read by mobile handheld readers and higher speed scanners. It is, however, prone to faults since multiple scanner are required during the entire transportation process. Another technology is a Radio Frequency Identification– RFID. It doesn’t need visual contact. It does not need any direct contact between the tag and the reader. It is available in different sizes. It provides different functionalities. It is easy to use and it has different costs associated with it. It can be active, passive, and semi-active. It works based on the radio frequency and it is generally rewritable. Another identification technology– a more recent one– is Near Field Communication– NFC.
It does not require visual contact or direct connection between tag and the reader and it does not emit radio frequency. In addition to identification, sensing is also very important. In 1999, a group of people led by Khan and Pister came up with the idea of a smart dust– what they called a network of tiny sensors, robots, or devices installed with wireless communications that can detect anything from light and temperature to vibration and et cetera. The idea was to have very small computing and sensing devices and to throw them in the environment like dust, and they start sensing, networking, and measuring. But they were not about sensors only.
They are talking about sensors with brains– sensors that can sense, think, and talk. Sensors are often cheap, unreliable, and imprecise. The extra advantages offered by having small brains on the sensors are allowing them to function autonomously, to execute business rules, bringing logic of the enterprise close to the point of action. It also enable the sensing devices to improve the quality of their readings, to remove false readings, or to calibrate, or to reduce the response time– all in all, to make the network more efficient, smarter, and more autonomous. But what can be sensed in the first place? It differs, depending on the good and product monitor or to transport.
Sensors are used to monitor quality of goods, like temperature, light, humidity, quality of the transport condition, like motion, or the exact location of the good, like GPS. Individual sensors provide valuable information, but their combination often give even more information. An example is high temperature only may indicate a warm and dry condition during the transport, but it may also indicate warm and wet condition, which may not be good for the product to be transported. Combining these two different sensors, meaning combining temperature and humidity, provides exact condition. The information can be also used to identify events or errors. Error is also known as anomaly, exception, faults, defects, or noise.
It may occur because of human errors, instrumental errors, mechanical faults, or change in the environments. Error in data reduces the quality of decision making and should be identified and immediately discarded. The process of accurate and timely identification of these errors is often called outlier detection. Detecting outliers when and where happen offer two main advantages– increasing quality of data, and consequently, increasing quality of the decision-making process, and reducing the amount of data to be transmitted. What is an event? Sensor values often have a normal pattern, which is followed on the normal circumstances. Consider a truck– a truck transporting roses, for instance.
The normal condition is to have this rock under certain temperature, but if the temperature rises to different effects– for instance, truck being long time under sun. This is an event. The variation from these normal patterns is an indication that something out of, ordinary or an event, has occurred. Out of ordinary in this case means often unwanted situations. These occurrences should be detected and further analysed to get to the bottom of what actually has happened. Sensor notes or sensing devices as standalone devices are limited and unreliable. Detecting errors and events require cooperation and communication between different sensing devices. Outlier detection and event detection require cooperation and communication between different sensing devices. This is where networking plays an important role.
Networking is the topic of the following lecture.