Answer To: Subject- project planning (ENRP20001) Topic- Design and analysis of hydraulic controlled sliding...
Amar answered on Apr 26 2021
Design and Analysis of Hydraulic / Automatic Controlled Sliding Door
Automatic form of entrance / exit control of doors is widely employed across public places like that of businesses, grocery stores, airports, transportation stations as well as wholesale departmental / retail stores for eliminating the needs of opening as well as closing actions in a manual manner. The contemporary sensor dependent control technologies for automatic doors encompass ultrasonic / radio, infrared, or else other wireless sensing approaches (Yang et al. 2013; Oladunmoye et al. 2014; Zhang et al. 2013). The infrared approach could be further be classified as active as well as passive approaches. In the active process, infrared signals are emitted by way of controller as well as captures signals that are reflected for the purposes of determining in case there are any objects closer to a door. The approach can be noted to be accurate as well as capable to identify the overall position as well as speed pertaining to the object and yet the high cost pertaining to the same have made it to be less popular. In contrast, passive approach detects the infrared signals that are radiated from people as well as is one of the methods that is used most widely on account of being low cost, effective and simple. In the other end, ultrasonic / radio method entails the emitting of radio / ultrasonic waves for the purposes of scanning the overall environment as well as analysing the signals returned for access control with respect to the door (Yang et al. 2013; Oladunmoye et al. 2014; Zhang et al. 2013).
Even though the techniques discussed have all been successful to detect the objects, they have been noted in not being capable for understanding specific intention as well as type of objects in relation to the door (Yang et al. 2013; Oladunmoye et al. 2014; Zhang et al. 2013). To take an example, passing pedestrians or else a puppy could trigger door in an accidental manner and thus cause false / unnecessary opening action. The frequent instances of these false action shall not just be annoying, but in addition lead to energy waste of air conditioning, as well as reduce the lifetime of the equipment. The same essentially makes it imperative in terms of a need for automatic form of door controlling system that is on the basis of detecting as well as undertaking intention analysis of individuals. This innovative system for door control shall be predominantly be on the basis of human detection as well as the intention assessment (Yang et al. 2013; Oladunmoye et al. 2014; Zhang et al. 2013). The initial part shall entail the detection of faces or else contour detection which will aid in identifying if the object detected is actually a person, whilst the latter shall include tracking of trajectory as well as statistical analysis with respect to the estimation of intention. The flowchart pertaining to control procedures in this regard is presented under the Figure 1 below.
Figure 1 – System Flowchart (Yang et al. 2013)
It can be noted that the detection of faces at reasonable levels of detection rates have been noted to be well developed in the manner of reviewing the literature in this regard as well as have closest relevance in terms of human characteristics, and hence can be noted to be effective for the identification of people (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016). In summary, once an individual gets detected within the specific region of interest (“ROI”), overall trajectory concerning the individual’s face could be suitably tracked as well as later on analysed by way of a statistical analyser for the purposes of calculating the relevant cumulative probability value, as well as the same shall be employed in being an estimation for the specific intention. In the other end, as depicted in the Figure 1 above for the procedure of door-closing, door shall be closed in case of none of the object shall get detected in passage over both the sides of door (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016). At a broad level, processes concerning the transition of state, detection of humans, intention assessments, as well as theoretical performance assessments shall all be crucial in this context.
The state diagram, as presented under Figure 2, represents the system for accessing control and the same entails opening as well as the closing actions. The process of door opening essentially identifies detect targets to be human by way of detecting the face / contours, as well as by calculating the intention probability for door access, referred as Paccess, by way of analysing the pertinent trajectory door access. Once this Paccess has been noted in being greater than that of threshold pre-set value, said door shall be opened in a suitable manner. In the other end, infrared sensors shall be added for making sure that no one is passing, or else stayed prior to door closing, as well as later activating close action while entrance / exit can be noted to be clear to be safe (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016).
Figure 2 – State Transition (Gunoz et al. 2018)
Over the recent times, techniques of human detection, specifically the ones which are implemented using the strategies of face detection, are noted in being applied successfully for many of the consumer products like that of smart phones, surveillance systems or else digital cameras, to detect people (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016). In one of such systems reviewed, even though both of the face detection as well as contour detection have been adopted to undertake human identification, face detection can be noted in being employed to be a primary solution as well as that contour detection is on the basis of head as well as shoulder shapes, and are left to be an optional solution with respect to application that have specialized form of privacy considerations (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016). Hence, the function of human detection can be noted to be represented using face detection rather than contour detection. The research studies focussed on face detection could be categorized in being of four different types: [1] knowledge-based, [2] feature-based, [3] template matching-based as well as [4] appearance-based methodologies (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016).
In the initial method, that is, knowledge-based, face shall be identified using the face features that are constructed using relative locations as well as distances relating to nose, eyes as well as mouth. The rate of detection shall be affected to a strong extent by way of orientation pertaining to the face. In the context of feature-based method, the features of face shall be constructed using statistics pertaining to locations as well as distances in-between facial organs (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016). Face shall be detected using matching of regions having similar set of features as well as the information of skin colour. The rate of detection using this approach shall be better as compared to knowledge-based approach, and yet the overall performance shall be easily influenced on account of conditions pertaining to lighting, shadows, and noises. The approach of template matching shall employ pre-determined facial templates to undertake matching. The face shall be identified after Euclidean distance in-between ROI as well as the internal template shall be lesser than that of specific threshold value. The said approach can be easier for being implement and yet the same is not noted to be versatile with respect to varied types in face geometry. With respect to the appearance-based approach, techniques of machine learning as well as statistical analysis shall be employed for the purposes of finding the optimal features pertaining to faces. Even though longer time for learning as well as greater number of samples shall be required with respect to training process, the stated methods possess higher rates of detection as well as lower time for processing, and shall be highly practical with respect to applications in real time and thus have high potentials for being adopted widely (Gunoz et al. 2018; Lachheb et al. 2016; Jing 2018; Lee and Lee 2016).
Appearance-based approach can be further classified as three different types: [1] linear / non-linear projection, [2] neural network, and [3] Gabor filter / Haar like features selection (Corpuz et al. 2015; Bhatt et al. 2018; Xie et al. 2013). For instance, principal component analyses (“PCA”), linear discriminate analyses (“LDA”), as well as independent component analyses (“ICA”) represent popular forms of projection-based methodologies using which features shall be extracted by way of face images being projected to the lower dimension feature spaces wherein highly significant form of feature vectors shall get selected. After the same, input sample vectors having distances, like for example Euclidean distance, that are shorter as compared to pre-determined threshold shall get determined to be the face image (Corpuz et al. 2015; Bhatt et al. 2018; Xie et al. 2013). In essence, performance pertaining to the stated type shall depend to a large extent over suitable distance measurements. At a contrasting context to this projection schema, neural network schema employs significant...