How to Write Research Paper Introduction in 03 easy steps

Introduction Section


Can you just think of starting your research paper directly from the methodology you used to solve your research problem?  You can not simply pour your methodology, ideas, and arguments without explaining what are you writing about.

Similar to a  novel or story the first step for writing a research paper introduction is to introduce the audience to the background(survey) and the major issue  (research problem)  different roles (methods)  and the theme around which the story revolves (the proposed method).

In the introduction section, you must clearly indicate the problem you want to solve. You must explain the necessity of your work, its significance and how your work may change the future course of research.  Your writing appeal should be such that it should hook readers to continue reading it.

The introduction does not have a strict word limit, unlike the abstract, but it should be as concise as possible. A good introduction explains how the research problem has been solved by various researchers (Literature survey)   and creates ‘leads’ to make the reader want to delve further into the research domain.

An introduction has three invisible sections namely

  1. History or background of the research
  2. Survey of the relevant papers
  3. Problem formulation and methodology used to solve the problem

1. History or Background of the Research

An introduction is the first paragraph of a  research paper. Start your introduction with short history. There can be two possibilities to start an introduction. One way to tell your readers is about a seminal paper, research funding, special event, or invention of an algorithm which spurred the development of the field under investigation.

For example, the introduction of Big Data gave a new impetus to the storage and analysis of huge data sets. Another option can be; the first paragraph of your introduction should be a historical narrative, from the very first research in the field to the current day with key statistics describing the development of the field. Here in both cases introduce the keywords of the field and describe what the various keywords mean.

2. Survey of the Relevant Papers

Provide an overview of existing thinking about and/or research into your research problem. Identify a gap, problem or unresolved issue in the existing knowledge/research that your research can fill or identify a research focus that will be useful.  Here your citations are crucial.  Try to survey the papers in which the authors are authoritative in your research domain.

3. Problem Formulation and Methodology used to Solve the Problem

The reader, by the end of the introduction, should know exactly what research issue you are trying to solve with your paper. State the intent of your study, including the research question and your unique methodology to solve the problem.

Describe important results that you have found or hope to find. The introduction leads the reader from a general research issue or problem to your specific area of research.  While specifying the methodology state the algorithm /the standard tool /technique which will be used for implementing the methodology.


An Optimized Fuzzy Based Short Term Object Motion Prediction for Real-Life Robot Navigation Environment

Short Term object motion prediction in a real-life Robot navigation environment refers to the prediction of the next instance position of a moving object based on the previous history of its motion. Living beings and vehicles characterize the dynamic environment and exhibit motion in various directions with different velocities.

In an unmanned Robot navigation system, the Robot has to acquire information about moving objects and predict their future positions for the next instance based on their previous history of motion in order to make efficient path planning[3][7].

The sensors available to read the position of the moving object should send accurate data in quick time succession and the Robot should process and generate the predicted position within a limited time as the validity of the result is very short.

Real-life data often suffer from inaccurate readings due to environmental constraints, sensors, the size of the objects and possible changes in the motion pattern of the moving objects. This needs the system to be Robust to handle these uncertainties and predict the next instance object position as accurate as possible.

Research literature has addressed solutions to the short-term object motion predictions with different methods such as Curve fitting or Regression methods [9] [22]Neural network based approaches [1][2][5], Hidden Markov stochastic models [23], Bayesian Occupancy Filters [6], Extended Kalman Filter[12][18] and Stochastic prediction model [21].

Based on the literature survey it is observed that i) The existing models lack flexibility in handling the uncertainties of real-life situations. ii) Probabilistic models sometimes fail to model real-life uncertainties. iii) The existing prediction techniques show poor response time due to their complex algorithmic structure. iv) Most of the approaches validate the results with simulated data or simple navigational environments.

The present work overcomes these difficulties with a simple solution for short-term motion prediction using the fuzzy inference method. It is assumed that the Robot is instantaneously stationary and observes the moving object through vision-based sensors.

The position of the moving object is sampled by the Robot at two definite time intervals. With the current sampling positions, the proposed Fuzzy predictor algorithm predicts the future object’s position in the following sampling duration.

The model is flexible as the navigation environment considered is fuzzy in nature. In the initial step, the Fuzzy rule base is generated by the simulator with different object motion patterns within the navigational environment.

Inconsistent and redundant rules are removed by defining directional space within navigational environment. Even though the fuzzy inference rules are relatively more in number, still the response time of the predictor is very small due to the short circuit evaluation of rules[13].

Min Max, Centre of Area and Mean Of Max defuzzification techniques are used to generate crisp output to identify the suitable defuzzification technique in terms of accuracy and quick response time for the current application. The algorithm is tested on both simulated and real-life benchmark data provided by the CAVIAR project[10]. The complete process of short-term motion prediction is represented in Fig1.

In fact, many research supervisors advise research scholars to finish the body of the paper as a draft and then work on writing the introduction. This method is sometimes advantageous as it will show a clear flow of research carried out and a small discussion on results in the introduction section which may arouse the interest of the reader to read the paper in depth.  However, with a well-defined outline, a researcher can start writing the introduction first. In this case, after completion of the paper writing researcher has to check the flow of the rest of the paper and the introduction section to keep the thread consistent throughout the paper.

Vijay Rajpurohit
Author: Vijay Rajpurohit