Sage Journals: Discover world-class research

Abstract

The cordial existence of robots and humans in the same workstation is an important aspect of the deployment of autonomous systems. Usually, robots are integrated with auxiliary sensors for monitoring the surroundings and tackling unexpected actions. The incorporation of LiDAR (light detection and ranging) and SLAM (simultaneous localization and mapping) has gained the researcher’s attention these days. The current study aims at developing a collision-free motion for an autonomous robot to ensure the least possible travel period at internodes. First, a SLAM-based algorithm for a static environment is proposed. A low-cost microcontroller Arduino UNO is used here along with LiDAR VL53L0X and is integrated with MATLAB to store and display the acquired data. The A* algorithm is used to navigate between the shortest paths on the obtained map. The robot is also integrated with radio frequency identification (RFID) sensors to automate the process. The proposed system would be useful for material handling, assistance, and cleaning purpose in large warehouses or households.

Keywords

LiDAR SLAM neural network autonomous vehicle inventory control and assistance robot

Introduction

Some of the areas that signify the use of smart and intelligent logistics are automating workflow and process automation, supply chain, robots, transport, black boxes, better demand prediction, live tracking, etc.¹ Robotics in logistics is a very broad topic. In fact, it concerns every area of the supply chain—that is from warehouse automation to robotics used in ports, and to drone deliveries.² Looking little bit deeper, narrowing down the subject to strictly warehouses or households, the robotics must be considered in such aspects as stated here. The first is the automation of the flow of goods (installation of pallet or container conveyors allows improvement in the flow of goods between different areas of the warehouse, and even between different facilities).^3–6 The second is automation of picking process (implementation of semi-automatic systems such as pick-by-light or devices for voice picking pick-by-voice, which support warehouse workers in the process of preparing orders, indicating, e.g. the number of goods to be picked) or order dispatching automation (installation of equipment sorting goods by carrier or equipping docks with a system for automatic loading and unloading of goods enhanced with the installation of devices sorting goods by carrier or equipping docks with a system for automatic loading and unloading of goods).^7–9 The robots currently available are suitable for large-scale operations, however, small-scale integration of these robots is not only costly but also impractical.¹⁰ Small-scale use of such robots can be characterized by less load carrying capacity and less computational strain at less expense, with technologies to integrate these robots into the internet use of such robots can be beneficial.^11–14

The robot proposed herein consists of an Arduino microcontroller and sensors used to monitor proximity detection and mapping purposes. The mobile robot considered here is wheeled indoor robot with LiDAR sensors mounted on top. Output of sensor is input for MATLAB which is connected with help of Bluetooth module. A* algorithm is used to navigate between the shortest paths on the obtained map. The robot is also integrated with radio frequency identification (RFID) sensors to automate the process. The objective is to design the prototype of the robot to achieve economics in the handling of raw material, work in progress, and finished products. The development of a conscience-friendly robot by integrating hardware and software components with ergonomics is proposed. An affordable proximity detection system for future development of autonomous robots and vehicles is advocated. It aims to increase in automation of small warehouses and businesses, productivity, and efficiency of warehouse operations. It assists rudimentary tasks to robots so workers can work without fatigue and would minimize the possibility of accidents. These robots can be used in any indoor environment due to their programming. The main focus is on small-scale warehouses where loading and unloading operations are quite frequent. This demand can be met by using a robot which consumes less power by traveling through the shortest path available. These robots can also be used in restaurants as a food delivery system. They can be used in hospitals for rudimentary tasks like transportation of goods and assistance services. They can be used in hotels for delivery and assistance services. Supermarkets are similar to inventories but with the more dynamic environment, with high computation these environments can also be successfully navigated to provide services to the customers such as item search and navigation assistance. Major contributions of this study are:

MATLAB-based controller was used to generate a map of indoor environment by using ToF (time of flight) sensor, which was integrated with A* algorithm to generate the shortest path.

The shortest path is followed by the robot by using pure pursuit controller while sensing the obstacles ahead by the same TOF sensor.

The robot is also integrated with RFID sensor to the sense the tag on objects placed on the robot and navigates autonomously from source to destination.

The robot is equipped with lead screw mechanism for raising and lowering the platform where objects are placed. The prototype is designed to carry and lift objects weighing 1 kg.

Design and calculations

The prototype is manufactured with mild steel angle of L-cross section for frame with plates of 2 mm thickness and wooden strips for mounting Arduino and other electronic components. The density of mild steel is 7850 kg/m³. The mild steel angle of a thickness 3 mm, height/width is 25 mm, mild steel plate of a thickness is 2 mm, wood strips of thickness is 8 mm, width is 15 mm. Figure 1 shows cross-section of a mild steel angle and its properties.

Figure 1.

Cross section of a mild steel angle and its properties.

Load calculations for MS plate

\begin{aligned} Cross - section of steel angle = 25 \times 5 + 20 \\ \times 5 mm = 225 m m^{2} \\ Total length of assembly = 80 \times 10 = 800 mm \\ Volume of frame = 225 * 800 + 2 \times 10 \times 20 \times 2 = 180, 800 m m^{3} \\ Density of steel = 8.05 g / c m^{3} \\ Weight of frame = 180.800 \times 8.05 = 1455.44 g \\ Weight of motor, Arduino, stepper motor and gear, \\ L i D A R, Battery = 1700 g \\ Total weight of robot = 1.4 + 1.7 = 3.1 kg \\ Load carrying capacity at 100 angle = 1 kg \\ Assume velocity of 15 m / \min = 0.25 m / s \\ Ignoring the friction of the surface \\ F = mg \sin ϕ = required motion \\ F = mg \sin ϕ + weight * acceleration \\ F = 4 \times 9.81 \times \sin 10 + 4 \times 0.25 m / s^{2} = 7.8139 \sim 8 N \\ Required force = 8 N \\ Assume driver wheel diameter = 0.04 \times 8 = 0.32 Nm \\ Required speed of rotation = (v \times 60) / (3.14 \times D) \\ Required speed of rotation = 60 rpm \\ Torque in kg - cm per motor = 3.2619 / 2 = 1.630 kg cm \end{aligned}

Calculations for Lifting Mechanism:

Torque required raising load is calculated as follows,

\begin{aligned} P_{r} = Effort required to raise the load (N) \\ P_{r} = \frac{W (\tan φ + \tan α)}{(1 - \tan φ . \tan α)} \\ W = Load (N) = kg = 9.81 N \\ φ = Friction Angle = \tan^{- 1} (μ) = \tan^{- 1} (0.4) = {21.80}^{\circ} \\ μ = Coefficient of friction = 0.4 (for Steel and Brass) \\ α = Helix Angle = \tan^{- 1} (\frac{l}{π . dm}) = \tan^{- 1} (\frac{12}{π \times 7}) = {28.62}^{\circ} \\ Here, φ < α, i . e . self - Hauling Condition of Lead Screw \\ Hence, P_{r} = \frac{W (\tan φ + \tan α)}{(1 - \tan φ . \tan α)} \\ P_{r} = \frac{9.81 (\tan 21.80 + \tan 28.62)}{(1 - \tan 21.80. \tan 28.62)} \\ P_{r} = 11.86667 N \end{aligned}

Again, torque for raise load is,

\begin{aligned} T_{r} = P_{r} \frac{dm}{2} = 11.86667 \times \frac{7}{2} \\ T_{r} = 41.5339 N . mm \end{aligned}

Table 1 shows a lead screw specifications.

Table 1.

Lead screw specifications.

Lead screw specifications	Dimensions
Screw material	Steel
Total length of screw	150 mm
Nominal/outer diameter	8 mm
Mean diameter $(d_{m})$	7 mm
Lead $(l)$	12 mm
Coefficient of friction $(μ)$	0.4
Nut material	Brass

Calculations for gear and pinion

Beam strength of gear

\begin{aligned} F_{b} = σ_{b} \times b \times m \times Y (Lewis Equation for Beam Strength) \\ Here, σ_{b} = \frac{Sut}{3} = \frac{75}{3} = 25 N / m m^{2} \\ b = 9 mm, m = 0.8 \\ Y = 0.484 + \frac{2.865}{z} (20^{\circ} Full depth Involute System) \\ = 0.45216 \\ Hence, F_{b} = 25 \times 9 \times 0.8 \times 0.45216 \\ F_{b} G = 81.39 N \end{aligned}

Beam strength of gear pair:

Beam strength of pinion

\begin{aligned} F_{b} = σ_{b} \times b \times m \times Y (Lewis Equation for Beam Strength) \\ Here, σ_{b} = \frac{Sut}{3} = \frac{440}{3} = 146.667 N / m m^{2} \\ b = 8 mm, m \times 0.8 \\ Y = 0.484 + \frac{2.865}{z} (20^{\circ} Full depth Involute System) = 0.41237 \\ Hence, F_{b} = 146.667 \times 8 \times 0.8 \times 0.41237, FbP = 387.083 N \\ Thus F_{b} G < F_{b} P \end{aligned}

Hence, gear is weaker in bending; beam strength of Pinion is safe.

Wear Strength of Gear Pair:

Wear strength is maximum tangential load the gear tooth can take without wear failure.

\begin{aligned} Fw = dp \times b \times Q \times K \\ dp = 32 mm, b = 8 mm \\ Q = \frac{2. z G}{z G + z P} = \frac{2 \times 90}{90 + 40} = 1.3846 \\ K = 0.16 (\frac{BHN}{100}) 2 BHN = 125 = 0.16 \times 1.252 = 0.25 \\ Fw = 32 \times 8 \times 1.3846 \times 0.25 = 88.61 N \end{aligned}

Hence, design is safe.

Table 2 shows a gear and pinion specifications.

Table 2.

Gear and pinion specifications.

Parameter	Gear	Pinion
Material	Nylon plastic	Mild steel
Pitch circle diameter (d)	72 mm	32 mm
No. of threads (z)	90	40
Face width (b)	9 mm	8 mm
Ultimate tensile strength (S_ut)	75 N/mm²	440 N/mm²

Selection of components

In this section, selection of motors, lifting mechanism, LiDAR & Arduino UNO is discussed.

Selection of motors

The motor of 4.5 kg-cm (0.44 Nm) was selected for drive operation and 4 kg-cm stepper motor for lifting mechanism, which is more than sufficient as per the design considerations. The power supply required to drive this motor is 6 V at 1 A. Hence for 1 hour of operation power supply of 6V and 3.5 Ah was selected.

LiDAR and Arduino UNO

LiDAR or light detection and ranging senses distance by pulse of a laser. Basically what a laser range finder does but it goes a step farther.¹⁵ The laser beam or laser pulse is swept across the target to be measured and the LiDAR device take multiply readings of the distance.¹⁶ This provides a profile cross section of the target. This repeats over and over again until a 3D picture or model can be constructed from the data. LiDAR is most often used on a moving device, such as a plane or satellite flying over a landscape.¹⁷ It can penetrate leaves in forests so as to get duel reflections from the ground and tree top which is very helpful in mapping the surface contours and determining tree height.¹⁸ LiDAR technologies use near infrared light to illuminate target surface, and then distance and reflectivity measure of the surface can be estimated from the reflected light.¹⁹ The key benefit of this technology is both elevation and intensity image can be obtained with high accuracy.²⁰ All reflected points are gathered to generate a 3D model of the illuminated object/surface. This is known as 3D point cloud. Therefore, 3D model of an entire area can be captured using this technology. Many useful products can be derived from LiDAR point cloud.²¹ This video can provide a preliminary idea about acquired LiDAR data. LiDAR is like a camera that takes a picture of distance to objects instead of the color of the object.²² For each sampling location the LiDAR scanner/camera records the distance to objects in the field of view creating 3D map of the volume of space scanned.²³ LiDAR can use different methods to achieve this. One uses a physical scanner that projects a pulsed laser beam in a fan shaped pattern that scans from left to right up to down and then repeats.²⁴ The laser fires an extremely short and intense pulse of light and the waits to see how long it takes to see a reflection of the light.²⁵ The longer it takes the farther away that piece of the world is from the scanner. If it does not see a reflection it assumes the world is very far away in that direction.²⁶ In a very short amount of time the laser has scanned a large area and returned a 3D map of the world around the scanner.²⁷ Figure 2 shows a circuit of VL53L0X and Arduino Uno. LiDAR VL53L0X consists of 940 nm laser VCSEL (vertical-cavity surface-emitting laser), VCSEL driver, and ranging sensor with advanced embedded micro controller.²⁸

LiDAR VL53L0X Measurements: 4.4 × 2.4 × 1.0 mm

LiDAR VL53L0X Range: Measures absolute range up to 2 m in indoor environment (reported range is independent of the target reflectance)

LiDAR VL53L0X is eye safe, has single power supply, I2C interface for device control and data transfer along with Xshutdown (reset) and interrupt GPIO.

Figure 2.

Circuit of VL53L0X and Arduino Uno.

LiDAR uses laser light at much higher frequencies than RF (300 MHz has a wavelength of 1 m, while UV and infrared is around 200–1000 nm).²⁹ This means that LiDAR works over a shorter range (because higher frequency signals attenuate faster), but it offers higher accuracy in resolving the detail of the object reflecting the laser. So it can be used for measurements such as 3D scanning of an object's surface, where the distance of the scan depends on the input power and efficiency of the system. Figure 3 shows a typical ranging performance in default mode of VL53L0X. As most of the computation was done on MATLAB, Arduino UNO was purely used for communication and control for which it was more than sufficient. Two drive motors, two stepper motors, two I2C interface/connections, and a Bluetooth module have been used in this investigation. The program uploaded consumed less than a quarter of Arduino's memory and the speed was sufficient for our application.

Figure 3.

A typical ranging performance in default mode of VL53L0X.

Use of open source platform such as ADXL335 accelerometer,³⁰ power sensors in face milling,³¹ microphones,³² gyroscopes,³³ thermocouples,³⁴ fluid flow sensors,³⁵ etc., integrated with microcontrollers such as Arduino,³⁶ Raspberry Pi,³⁷ etc., has been increased in design of such systems.^38,39 In addition to this, use of metal matrix composites for making body of the structure has been increased which makes the skeleton robust.⁴⁰ Application of ML-based techniques for decision-making in condition monitoring is also becoming popular day by day due to its intelligent nature.⁴¹

Lifting mechanism

The lifting mechanism used here is consists of a lead screw coupled with a rotating gear and a nut coupled to a square shaft which is used to lift the load up and down according to the rotation of gear which is rotated with the help of pinion, coupled to a stepper motor. The lifting capacity is limited to 1 kg for the purpose of prototyping.

Design of algorithms

In this section, design of algorithm is presented including A* algorithm, algorithm for RFID, SLAM algorithm for navigation and pure pursuit algorithm.

A* algorithm

A* is the best-first search, or informed-search code, which is formulated via weights: commencing from a specific node, it targets to discover a track to the certain aim node taking the lowest cost.⁴² Normal A* search initializes from a start state, which is prioritized by the sum of distance and a heuristic to the goal.⁴³

When the frontier reaches a goal state, an optimal path has been found. Bidirectional A* has a pair of search frontiers initialized from both the start and goal states.⁴⁴ Each node on the frontiers can be described either by distance from the start + heuristic to the goal, or by distance to the goal + heuristic to the start, whichever is appropriate.⁴⁵ When the frontiers meet, an optimal path has been found. But, it might face multiple goal states. Or, it might be that there is not any distinction between start and goal states—there is just a need of making a connection between a pair of nodes in a graph, for example.⁴⁶ Therefore, bidirectional search can be generalized to N-directional search, where the frontier is prioritized by:

distance from the start (or goal), plus the minimum of the heuristic to any of the goal states, OR

distance to the goal, plus the minimum of the heuristic from any of the start states⁴⁷

Computing the heuristic multiple times may eventually become too costly to be practical. Or, adding a lot of goal states to the search may negate some of the benefit. But “tri-directional search” can be practical and advantageous, depending on the problem domain.⁴⁸ Figure 4 shows a pseudo code for A* algorithm.

Figure 4.

Flow chart for A* algorithm.

RFID

RFID stands for radio frequency identification and as the name says the RFID technology works by identification of the particular frequency. Usually, these RFID consists of two parts, the transmitter and receiver. The receiver would be an RFID module whereas the transmitter is basically a card or a small key tag which is capable of producing a particular frequency.⁴⁹ No two RFID tags will produce the same frequency. Coming to the structure of the module, it consists of a microchip which can recognize the RFID frequency up to a certain range. Inside the transmitter, there will be a complete circuit. The power for the transmission of the radio frequency waves comes from a tiny capacitor which will be hooked up inside the circuit.⁵⁰ As soon as you hold the tag near the module, the electromagnetic field produced by the module charges the capacitor which acts as a power source for the tag, that is the transmitter. RFID, a popular radio frequency and identification technology, is used by various brands to monitor their stocks in warehouses.⁵¹ The RFID system is similar to a bar code scanning system that you can find in any supermarket. One of the best benefits of using an RFID system is they are quite precise when compared to any other technology. RFID is composed of two major elements. The tag is actually a sophisticated little (they can be grain-of-rice sized) device, that contains a loop antenna, a microchip, and sometimes an external capacitor.⁵² When the properly-tuned reader coil comes near it, some of the emitted radio signal from the reader is absorbed by the tag's antenna, and it uses some of that power to operate the microchip, which then returns a digital code using nearly the same radio frequency. The code can be short (as in 64-bits) or very long (up to several thousand bytes). However, the longer codes require more time and more power, so usually the designers choose the shortest code they can for the application.⁵³ RFID or radio frequency identification has many uses such as identifying and tracking items in retail stores. It is used to check inventory levels and also to curb shoplifting and theft by employees, in airports for customer luggage tracking, on large construction sites for tracking workers. It can even record slip and fall accidents, and workers can be tracked continuously. There are many more uses for RFID like child tracking etc. As the price of tags has drastically decreased since inception, more people are using this inexpensive method of control.⁵⁴

SLAM algorithm for navigation

The full name of SLAM is simultaneous localization and mapping, which translates to instant localization and map construction.⁵⁵ There are two key words here: localization and map construction, which means that the robot will determine its position in an unknown environment while building a map, and finally output a map like this.⁵⁶ SLAM used herein was the part of seeded toolbox provided in MATLAB which uses pose graphs, that is remembering few characteristics of the previously visited site. This technique is called bundle adjustment. Therefore when the same site is visited again the algorithm detects the site and closes the loop.⁵⁷ To implement SLAM one needs an odometry sensor and a LiDAR sensor. The odometer gives an idea of the position of your robot on the 2D map. The LiDAR sensor will helps to build the map during the learning phase. Also one can use an Xbox Kinect sensor to build a 3D map. But then there is need of a high graphics computer to process that map.⁵⁸ SLAM is used for mapping an unknown terrain. It also locates itself in the map it created. This is important for dynamic environment as the surrounding is constantly changing. This is why it was used for DARPA Grand Challenge in 2005 where autonomous vehicle competes. SLAM was selected herein for the application of inventory control as the environment is less dynamic and thereby eliminating the use of more data points.⁵⁹

Pure pursuit algorithm

It is a tracking code based on estimation of the curvature through which vehicle moves from source to its destination.⁶⁰ The main task is to select an aim and calculates the curvature.⁶¹ The input to the pure pursuit algorithm is a variable called LookAhead distance along with the path generated for the robot with A* algorithm as input.⁶² As LookAhead distance increases so does the curvature of path, the robot follows. This curvature is helpful for smooth motion of the robot as it avoids instantaneous turns from the shortest path generated with A* algorithm.⁶³ The drawback of using pure pursuit is that, as the LookAhead increases there is a possibility that the goal co-ordinates are crossed during transition. To compensate for it is required to provide a goal range instead of goal co-ordinates, that is the range is which the robot is said to have reached the goal.^63,64

Prototype development, testing and results

Figure 5 shows a prototype LiDAR operated inventory control and assistance robot. Figure 6 shows A* Algorithm running to find shortest path. Figure 7 shows a pure pursuit controller with robot (in red). The time required for scanning of 5 scans was 6 minutes. As the path is simple and short the time required for A* was 15 seconds but it may increase to 28 minutes for much complicated paths. The robot moves according to the path described by pure pursuit controller and requires user input if any obstacle is detected. LiDAR is the visible or near-visible equivalent of radar. So it is a system that is use for detecting objects (present or not hypothesis), ranging (determining the distance to an object) and tracking these objects. This is done with EM waves that are transmitted coherently to the receiver such that Doppler information can be extracted. In the case of LiDAR, it is using wavelengths in the nano-meter range, whereas radar uses micrometer down to centimeter range. Both terms, radar and LiDAR, describe systems used for ranging and direction finding of objects, but with different technology and different purposes.

Figure 5.

Prototype LiDAR operated inventory control and assistance robot.

Figure 6.

A* algorithm running to find shortest path.

Figure 7.

Pure pursuit controller with robot (in red).

RF waves transmit over long distances, and are reflected by large metal objects, which is why it was originally used in the military to detect ships and aircraft. However, there's not a lot of accurate detail to extract from these signals, so ranging and direction finding are always approximate. LiDAR is a newer technology (1960s, whereas radar was developed in the 1890s), with the name being a portmanteau of “light” and “radar.” The interpretation as an acronym supposedly came later. With the use of RGB-D (camera and depth) sensors it is apparent that the robot will be able to detect not just objects and human in its vicinity but also be able to predict their motion and thereby enhance safety and become more conscious-friendly. This can be implemented using YOLO (you only look once) like algorithms to detect objects which uses convolutional neural networks (CNN).

Conclusion

The design of the LiDAR system for proximity detection is incorporated for the development of an anti-collision system for vehicles/robots. This system is capable of sensing proximity up to 1.5 m with an accuracy of ±5 mm. This system is useful for indoor robots and the future development of autonomous and semi-autonomous Vehicles. It is a cost-effective solution with low computational needs. The time required for operation is less than 7 minutes for simple routes but may increase to 20 minutes for scanning and finding the shortest routes for long distances. The use of a servo control system is unreliable and needs to be replaced with more accurate control systems. Loop closure is a major problem encountered by SLAM and needs to replace with much more computational algorithms. For dynamic environments use machine learning algorithms like neural networks can be used to navigate the robot autonomously. The use of a powerful microcontroller like Raspberry Pi can eliminate the need for MATLAB as the programs can be directly run on the mentioned platform. The use of a powerful microcontroller can also eliminate the need for Bluetooth technology as the range of such modules is no more than 100 feet. Integration of Wi-Fi modules can not only connect the robot to the internet for real-time data analysis but also help to monitor the condition of the robot. The use of active RFID technology can help increase the range from a few millimeters to a few meters which are a considerable advantage over the traditional barcode or QR scanners. Linear actuators with low losses of frictional energy can save a lot of electric energy thereby reducing the number of charging cycles and increasing the battery life of the robot. The use of a Li-ion battery is not only beneficial because of less charging time but also less harmful to the environment. With the use of SLAM techniques like visual SLAM which uses RGB camera along with depth (TOF) sensor one can generate 3D point cloud. Also inertial measurement unit (IMU) sensors can be used to provide better odometry readings. Incorporation of machine learning techniques is a must for recognizing the objects in the proximity of robot. Along with this robot should also be able to sense hazardous situations (e.g., fire) with smoke sensors and alert the people around.

Footnotes

Acknowledgements

The manuscript was accepted and presented at 2nd International Conference and Exposition on Advances in Mechanical Engineering (ICAME-2022) held during 23–25 June 2022 at College of Engineering Pune (COEP), Maharashtra, India.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Abhishek D Patange

References

Issaoui

Khiat

Bahnasse

, et al. Toward smart logistics: engineering insights and emerging trends. Arch Comput Methods Eng 2021; 28: 3183–3210.

Woźniakowski

Zmarzłowski

Nowakowska

. Automation and innovations in logistic processes of electronic commerce. Inform Syst Manage 2018; 7: 72–82.

Dudziak

Szymlet

. Flow of goods in the warehouse management systems-problem analysis. Acta Universitatis Nicolai Copernici Zarządzanie 2020; 47: 27–34.

Wurman

D'Andrea

Mountz

. Coordinating hundreds of cooperative, autonomous vehicles in warehouses. AI Mag 2008; 29: 9–9.

Van den Berg

Zijm

. Models for warehouse management: classification and examples. Int J Prod Econ 1999; 59: 519–528.

Hompel

Schmidt

. Warehouse management: automation and organisation of warehouse and order picking systems. Berlin, Heidelberg: Springer Science & Business Media, 2006.

Kluska

. Support the decision processes in management through the automatic generation of simulation models of warehouses. Eur Res Stud 2021; 24: 379–398.

Karásek

. An overview of warehouse optimization. Int J Adv Telecommun Electrotech Signals Syst 2013; 2: 111–117.

Lewczuk

Żak

. Supportive role of dynamic order picking area in distribution warehouse. J Traffic Logist Eng 2013; 1: 37–40.

10.

Pearce

Mutlu

Shah

, et al. Optimizing makespan and ergonomics in integrating collaborative robots into manufacturing processes. IEEE Trans Autom Sci Eng 2018; 15: 1772–1784.

11.

Paprotny

Bergbreiter

Small-scale robotics: an introduction. In Workshop at the IEEE international conference on robotics and automation. Springer, Berlin/Heidelberg, 1–15, 2013

12.

Hines

Petersen

Lum

, et al. Soft actuators for small-scale robotics. Adv Mater 2017; 29: 1603483.

13.

Chen

Jang

Ahmed

, et al. Small-scale machines driven by external power sources. Adv Mater 2018; 30: 1705061.

14.

Pieskä

Kaarela

Mäkelä

Simulation and programming experiences of collaborative robots for small-scale manufacturing. In 2018 2nd international symposium on small-scale intelligent manufacturing systems (SIMS). 2018, April. (pp. 1–4). IEEE.

15.

Cheng

Wang

GY.

Mobile robot navigation based on LIDAR. In 2018 Chinese control and decision conference (CCDC). 2018, June. pp. 1243–1246. IEEE.

16.

Elrayes

Ali

Zakaria

, et al. Smart airport foreign object debris detection rover using LiDAR technology. Internet Things 2019; 5: 1–11.

17.

Keong

Shukor

SAA

Rahim

NA.

Development of a mobile platform with IoT for LIDAR. In J Phys Conf Ser 2021; 2107: 012042). IOP Publishing.

18.

Segireddy

Koneru

Wireless IoT-based intrusion detection using LIDAR in the context of intelligent border surveillance system. In Smart intelligent computing and applications. Singapore: Springer, 2020. 455–463.

19.

Pawar

Teli

Shetye

, et al. Blind-spot monitoring system using LiDAR. J Inst Eng (India): Series C 2022; 103: 1–12.

20.

Athavale

Ram

Nair

BB.

Low cost solution for 3D mapping of environment using 1D LIDAR for autonomous navigation. In IOP Conf Ser: Mater Sci Eng 2019; 561: 012104. IOP Publishing.

21.

Phukan

Sinha

, et al. Information encoding, gap detection and analysis from 2D LiDAR data on android environment. In Advanced computing and intelligent engineering. Singapore: Springer. 2020. 525–536.

22.

Fakhri

Mardiati

Mulyana , et al. Prototype design for object coordinate detection using RP LIDAR concept. In 2020 6th international conference on wireless and telematics (ICWT) 2020, September. (pp. 1–6). IEEE.

23.

Liu

. Airborne LiDAR for DEM generation: some critical issues. Prog Phys Geogr 2008; 32: 31–49.

24.

Jaboyedoff

Oppikofer

Abellán

, et al. Use of LIDAR in landslide investigations: a review. Nat Hazards 2012; 61(1): 5–28.

25.

Dong

Chen

. LiDAR remote sensing and applications. Boca Raton: CRC Press, 2017.

26.

Lim

Treitz

Wulder

, et al. LiDAR remote sensing of forest structure. Prog Phys Geogr 2003; 27: 88–106.

27.

Levoy

Pulli

Curless

, et al. The digital Michelangelo project: 3D scanning of large statues. In Proceedings of the 27th annual conference on computer graphics and interactive techniques 2000, July: 131–144.

28.

Prasetya

DSA

Arseno

Pamukti , et al. Experimental analysis of vehicle-to-vehicle communication using light detection and ranging (LIDAR) for detection and data transmission. In 2021 International conference on artificial intelligence and mechatronics systems (AIMS). 2021, April. (pp. 1–6). IEEE.

29.

Fermann

Hartl

. Ultrafast fibre lasers. Nat Photonics 2013; 7: 868–874.

30.

Khairnar

Patange

Pardeshi

, et al. Supervision of carbide tool condition by training of vibration-based statistical model using boosted trees ensemble. Int J Performabil Eng 2021; 17(2): 229–240.

31.

Tote

Pardeshi

Patange

. Automatic number plate detection using TensorFlow in Indian scenario: an optical character recognition approach. Mater Today: Proc 2022.

32.

Pardeshi

Pujari

Khairnar

, et al. Design and analysis of nylon composite wheel hub for electric kart. In AIP conference proceedings 2022, October. 2653, 020009). AIP Publishing LLC.

33.

Shewale

Mulik

S. S.

Deshmukh

S. P.

, et al. Novel machine health monitoring system. In Proceedings of the 2nd international conference on data engineering and communication technology. Singapore: Springer. 2019. 461–468.

34.

Nalavade

Patange

Prabhune

, et al. Development of 12 channel temperature acquisition system for heat exchanger using MAX6675 and arduino interface. In Innovative design, analysis and development practices in aerospace and automotive engineering (I-DAD 2018). Singapore: Springer, . 2019. pp. 119–125.

35.

Mulik

Patange

Jegadeeshwaran

, et al. Development and experimental assessment of a fluid flow monitoring system using flow sensor and Arduino interface. In Innovative design, analysis and development practices in aerospace and automotive engineering. Singapore: Springer, 2021. pp. 115–122.

36.

Patange

Jegadeeshwaran

Bajaj

, et al. Application of machine learning for tool condition monitoring in turning. Sound Vibr 2022; 56: 127–145.

37.

Patil

Pardeshi

Pradhan

, et al. Cutting tool condition monitoring using a deep learning-based artificial neural network. Int J Performabil Eng 2022; 18(1): 37–46.

38.

Khade

Patange

Pardeshi

, et al. Design of bagged tree ensemble for carbide coated inserts fault diagnosis. Mater Today: Proc 2021; 46: 1283–1289.

39.

Bajaj

Patange

Jegadeeshwaran

, et al. A Bayesian optimized discriminant analysis model for condition monitoring of face milling cutter using vibration datasets. J Nondestruct Eval Diagn Prognost Eng Syst 2022; 5.

40.

Jatakar

Mulgund

Patange

, et al. Multi-point face milling tool condition monitoring through vibration spectrogram and LSTM-autoencoder. Int J Performabil Eng 2022; 18(8): 570–579.

41.

Deo

. Data imputation and comparison of custom ensemble models with existing libraries like XGBoost, Scikit learn, etc. for predictive equipment failure. arXiv preprint arXiv:2111.10088 . 2021.

42.

Duchoň

Babinec

Kajan

, et al. Path planning with modified a star algorithm for a mobile robot. Proc Eng 2014; 96: 59–69.

43.

Hoffmann

Nebel

. The FF planning system: fast plan generation through heuristic search. J Artif Intell Res 2001; 14: 253–302.

44.

Cheng

Saxena

Likhachev

Bidirectional heuristic search for motion planning with an extend operator. In 2019 IEEE/RSJ international conference on intelligent robots and systems (IROS). 2019, November. (pp. 7425–7430). IEEE.

45.

Cserna

Doyle

Ramsdell , et al. Avoiding dead ends in real-time heuristic search. In Proceedings of the AAAI Conference on Artificial Intelligence. 2018, April. 32.

46.

Sanders

Schultes

Highway hierarchies hasten exact shortest path queries. In European symposium on algorithms. Berlin/Heidelberg: Springer, 2005, October. pp. 568–579.

47.

Felner

Moldenhauer

Sturtevant , et al. Single-frontier bidirectional search. In Proceedings of the AAAI conference on artificial intelligence . 2010, July. 24, pp. 59–64.

48.

Yoosefzadeh

Tareghian

Farahi

. Tri-directional scheduling scheme: theory and computation. J Math Modell Algor 2010; 9: 357–373.

49.

Yamano

Tanaka

Hirayama

, et al. Self-localization of mobile robots with RFID system by using support vector machine. In 2004 IEEE/RSJ international conference on intelligent robots and systems (IROS)(IEEE cat. No. 04CH37566) 2004, September. 4, pp. 3756–3761). IEEE.

50.

Xiao-Long

Chun-Fu

Guo-Dong

, et al. A robot navigation method based on RFID and QR code in the warehouse. In 2017 Chinese automation congress (CAC) . 2017, October. pp. 7837–7840. IEEE.

51.

Mezzanotte

Palazzi

Alimenti

, et al. Innovative RFID sensors for internet of things applications. IEEE J Microw 2021; 1: 55–65.

52.

Wen

Zhu

. Swarm robotics control and communications: imminent challenges for next generation smart logistics. IEEE Commun Mag 2018; 56: 102–107.

53.

Morenza-Cinos

Casamayor-Pujol

Pous

. Stock visibility for retail using an RFID robot. Int J Phys Distrib Logist Manage 2019; 49(10): 1020–1042.

54.

Makris

Michalos

Chryssolouris

. RFID driven robotic assembly for random mix manufacturing. Robot Comput Integr Manuf 2012; 28: 359–365.

55.

Durrant-Whyte

Bailey

. Simultaneous localization and mapping: part I. IEEE Robot Autom Mag 2006; 13: 99–110.

56.

Fuentes-Pacheco

Ruiz-Ascencio

Rendón-Mancha

. Visual simultaneous localization and mapping: a survey. Artif Intell Rev 2015; 43: 55–81.

57.

Bailey

Durrant-Whyte

. Simultaneous localization and mapping (SLAM): part II. IEEE Robot Autom Mag 2006; 13: 108–117.

58.

Stachniss

Leonard

Thrun

Simultaneous localization and mapping. In springer handbook of robotics. Springer, Cham. 2016. 1153–1176.

59.

Frese

. A discussion of simultaneous localization and mapping. Auton Robots 2006; 20: 25–42.

60.

Coulter

. Implementation of the pure pursuit path tracking algorithm. Pittsburgh: Carnegie-Mellon UNIV Pittsburgh PA Robotics INST, 1992.

61.

Samuel

Hussein

Mohamad

. A review of some pure-pursuit based path tracking techniques for control of autonomous vehicle. Int J Comput Appl 2016; 135: 35–38.

62.

Wang

Hsu

TS.

The improved pure pursuit algorithm for autonomous driving advanced system. In 2017 IEEE 10th international workshop on computational intelligence and applications (IWCIA). 2017, November. pp. 33–38. IEEE.

63.

Wang

Chen

, et al. Trajectory tracking and speed control of cleaning vehicle based on improved pure pursuit algorithm. In 2019 Chinese control conference (CCC). 2019, July. pp. 4348–4353. IEEE.

64.

Wang

Gong

, et al. Development and evaluation of two learning-based personalized driver models for pure pursuit path-tracking behaviors. In 2018 IEEE intelligent vehicles symposium (IV). 2018, June. pp. 79–84. IEEE.

65.

Thrun

Simultaneous localization and mapping. In robotics and cognitive approaches to spatial mapping. Springer, Berlin, Heidelberg. 2007. pp. 13–41.

66.

Park

Lee

Han

WY.

Development of lateral control system for autonomous vehicle based on adaptive pure pursuit algorithm. In 2014 14th international conference on control, automation and systems (ICCAS 2014).2014, October. pp. 1443–1447. IEEE.

Development of LiDAR operated inventory control and assistance robot

Abstract

Keywords

Introduction

Design and calculations

Load calculations for MS plate

Calculations for gear and pinion

Selection of components

Selection of motors

LiDAR and Arduino UNO

Lifting mechanism

Design of algorithms

A* algorithm

RFID

SLAM algorithm for navigation

Pure pursuit algorithm

Prototype development, testing and results

Conclusion

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

ORCID iD

References