Analysis of Extra-Planet Searching Approaches: Radial Velocity,

Transit and AI Algorithm Detection

Yunjie Ruan

Shanghai Hongkou Secondary School Affiliated to Shanghai Normal University, Shanghai, China

Keywords: Transit, Radial Velocity, AI Detection, Habitability, Extraterrestrial Life.

Abstract: With the advancement of human exploration of the universe, the detection of exoplanets has become one of

the core areas in astronomy. The search for exoplanets and potential signs of life holds critical significance

for understanding galactic systems across the cosmos and the origins of life. This study analyses the current

mainstream detection methods, including the transit method, radial velocity method, and Artificial

Intelligence-based (AI-based) detection algorithms. It elaborates on their principles, technical advantages,

limitations, and major research achievements to date. According to the analysis, the transit method is highly

sensitive to the orbital inclination of planetary systems. While it can accurately determine a planet's radius, it

is unable to provide continuous long-term observations of radial velocity. The radial velocity method can

estimate planetary mass but is primarily effective for detecting massive planets in close proximity to their

host stars. Due to significant interference from stellar activity, this method poses considerable challenges in

practice. In contrast, Artificial Intelligence (AI) algorithms leverage deep learning to integrate multiple data

sources, enabling improved identification and analysis, thereby substantially enhancing detection accuracy.

These findings contribute to the development of an intelligent astronomical observation system that integrates

multiple technologies, thereby advancing scientific research in exoplanet detection.

1 INTRODUCTION

In the research history of exploring exoplanets, the

earliest was in 1988 when Canadian astronomers

discovered a Jupiter-like planet orbiting the star

gamma Cephei A, but it was not confirmed at that

time. Later in 1995, there was a major breakthrough

when Swiss astronomers discovered 51 Pegasi b,

which is a widely recognized exoplanet and also

opened a new era of exoplanet exploration (Mayor &

Queloz, 1995). Subsequently, various methods such

as the transit method, radial velocity method, and

microlensing method have continued to develop.

Researchers worldwide have utilized advanced

detection equipment and employed highly skilled

personnel to conduct in-depth studies.

Up to now, people have discovered more than five

thousand exoplanets, and there are still several

thousand candidates waiting to be confirmed.

Researching exoplanets can not only improve the

entire theoretical system of planetary formation,

allowing people to understand the formation

conditions and evolutionary processes of different

planetary systems, but also explore planets within the

habitable zone of stars, searching for various signs of

life (Charles, 2020). All of these can deepen people's

understanding of the universe and satisfy their strong

curiosity about the universe from ancient times to the

present (Wang, 2022). The full text focuses on three

detection methods: the transit method, radial velocity

method, and AI algorithm method. Incorporating

recent remarkable technological advancements, this

paper will propose an integrated approach combining

these three research methodologies.

In 2022, the total number of confirmed

exoplanetary systems reached 5,000. In September of

that year, astronomers announced the discovery of a

new type of exoplanet orbiting a relatively close red

dwarf, observed by National Aeronautics and Space

Administration (NASA) Transiting Exoplanet Survey

Satellite (TESS). Additionally, a scorching exoplanet

named TOI-1075b was discovered, one of the largest

"super-Earths" found to date with a surface

temperature of 1,050°C. Astronomers also detected

metallic barium in the atmospheres of two hot

exoplanets, WASP-76b and WASP-121b, the

heaviest element ever identified in exoplanetary

atmospheres (Polanski, et al., 2024).

364

Ruan, Y.

Analysis of Extra-Planet Searching Approaches: Radial Velocity, Transit and AI Algorithm Detection.

DOI: 10.5220/0013825900004708

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Innovations in Applied Mathematics, Physics, and Astronomy (IAMPA 2025), pages 364-368

ISBN: 978-989-758-774-0

In 2023, scientists made significant discoveries.

The James Webb Space Telescope confirmed the

presence of heavy elements like carbon and oxygen

in the atmosphere of the distant exoplanet HD

149026b, differing from the hydrogen- and helium-

dominated atmospheres of gas giants in the Solar

System. NASA's TESS discovered a gas giant named

TOI-4600c with an orbital period of 482.82 days, the

longest "year" among TESS-discovered planets, and

a surface temperature of -78°C. A rare planetary

system, HD 110067, was also found, featuring six

tightly orbiting "sub-Neptunes."

In 2024, a research team led by Professor Dong

Subo from the Department of Astronomy at Peking

University's School of Physics reported a new

exoplanet, Gaia22dkvLb, through follow-up

photometric observations of a microlensing event

alerted by the Gaia satellite. This planet has a mass

0.6 times that of Jupiter and orbits its host star at

approximately 1.4 astronomical units. The host star is

the brightest ever found for a microlensing-detected

planet, offering potential for measuring the planet’s

orbital eccentricity and other parameters via high-

precision radial velocity methods, as well as

searching for additional planets within its orbit to

study the overall architecture of the planetary system

(Rojas-Ayala, 2023).

In subsequent sections, this study will introduce

each detection method along with its fundamental

principles, describe the relevant instrumentation and

equipment used and present significant detection

results from recent years. Through comparative

analysis of outcomes from different methods, the

study will synthesize these three approaches before

concluding with future research prospects.

2 DESCRIPTIONS OF PLANET

SEARCHING

To detect exoplanets, it is essential to determine their

fundamental planetary parameters, including mass,

diameter, and orbital period. Equally crucial is the

investigation of atmospheric characteristics to assess

potential biosignatures, e.g., the presence of water

(H₂O), carbon monoxide (CO), and carbon dioxide

(CO₂) along with the identification of various

elemental compositions.

The specific process involves first determining

whether planets exist around the star, then measuring

the planet's orbital period, semi-major axis,

eccentricity, etc. This study uses these known

parameters to evaluate the physical properties of

planets, estimate their mass, radius, density, etc.

Subsequently, researchers observe whether

atmospheric components exist and search for signs of

life. Finally, they evaluate whether the planet lies

within the star's habitable zone.

Furthermore, evaluating planetary habitability

constitutes a critical component of exoplanet research.

The detection of liquid water and elements essential

for biological metabolism or replication provides key

evidence for potential extraterrestrial life. Current

mainstream exoplanet detection methodologies

include, i.e., Transit method, Radial velocity method

(Deng & Tang, 2024), Astrometry, Gravitational

microlensing, Direct imaging (Chauvin, 2023).

3 RADIAL VELOCITY

The radial velocity method is also called the wobble

method. Its principle is to detect exoplanets by

measuring changes in the star's motion speed toward

or away from Earth. Academically, it relies on

gravitational interaction planets orbiting stars form a

binary system, and while planets orbit the center of

mass in large orbits with wide ranges, the star actually

exhibits small wobbles.

Furthermore, when the star moves away from the

planet, its spectral lines shift toward the red end (i.e.,

wavelength becomes longer); when the star moves

toward the planet, its spectral lines shift toward the

blue end (i.e., wavelength becomes shorter). Through

long-term monitoring of these stellar spectral shifts,

one can calculate the periodic amplitude of velocity

changes. This allows us to deduce the planet's

existence and estimate parameters such as its mass or

orbital radius.

Otherwise, for the radial velocity method, there

are two types of high-precision detection instruments.

One is the High Accuracy Radial Velocity Planet

Searcher (HARPS), and the other is the High-

Resolution Echelle Spectrometer (HIRES). Through

complex mathematical calculations based on Kepler's

laws and Newton's law of gravitation, researchers can

conclude that:

𝑣 ∝

𝐺𝑀𝑝sin𝑖

𝑃

√

𝐺𝑀

∗

(1)

where i is one of the planet's orbital inclinations and

G is the gravitational constant.

Swiss astronomers detected a star in Pegasus

using the radial velocity method and observed an

oscillation every 4.23 days. Through data points and

fundamental principles, they calculated that there is a

planet close to Jupiter's size located 4 million miles

away from this star (Mayor, & Queloz, 1995).

Analysis of Extra-Planet Searching Approaches: Radial Velocity, Transit and AI Algorithm Detection

365

In recent years, while there are higher-precision

detection methods available, as one of the earlier

methods that has successfully detected multiple

exoplanets, it is still widely used and has its own

advantages and limitations. It can quickly detect low-

mass planets, but due to unknown orbital period data,

scientists need to conduct long-term observations.

Moreover, it is easily affected by stellar activity

interference and cannot directly measure planetary

radius. This method is one of the classical means for

detecting exoplanets and is often used in combination

with the transit method.

4 TRANSITS

The transit method, as one of the most classical

methods, has recently made significant discoveries.

Its detection principle is that when a planet passes

through the disk of its parent star, a slight decrease in

the star's visual brightness can be observed. However,

the degree of dimming depends on the size of the

planet itself and the size of the star. These brightness

changes cannot be observed by the naked human eye,

thus requiring long-term monitoring to analyze data

such as the planet's size. For the transit method, there

are two representative high-precision detection

instruments: the Kepler Space Telescope and the

TESS.

When a planetary transit occurs, telescopes can

also analyze the starlight passing through the planet's

atmosphere, which allows determination of the

atmospheric composition. This enables the search for

gases related to life. On April 17, 2024, a team of

astronomers from the University of Cambridge

published in The Astrophysical Journal Letters that

they detected Dimethyl Sulfide(DMS) and Dimethyl

Disulfide (DMDS) in the atmosphere of exoplanet

K2-18b, compounds only produced in Earth's

biological processes, also known as chemical

fingerprints. This planet is located in the habitable

zone of a red dwarf star 124 light-years away in the

direction of Leo (Madhusudhan, et al., 2025). A

typical measurement for the gas components is shown

in Fig. 1. This currently represents the strongest

evidence for extraterrestrial life, but this is not an

official announcement and requires more

observational verification.

Figure 1: A typical analysis of the gas components (Madhusudhan, et al., 2025).

5 AI ALGORITHM DETECTION

In recent years, the AI team has gradually grown

stronger. In the field of astronomy, the AI-driven

intelligent exoplanet detection team has also been

gradually introduced. For AI intelligent detection, its

principle leans more toward letting AI conduct data

learning, allowing it to learn from large amounts of

known exoplanet data, then through deep learning

algorithms master various phenomena and known

methods to deduce periodic changes in stellar

brightness, thereby calculating the probability of

whether certain specific characteristic patterns are

caused by planets. These data can all be provided to

astronomers for further research.

Currently, the Shanghai Astronomical

Observatory, with Professor Ge Jian leading an

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy

366

international team, has innovated a deep algorithm

combining Graphics Processing Unit (GPU) phase

folding and convolutional neural networks. From the

stellar photometric data released by the Kepler

telescope in 2017, they discovered five ultra-short-

period planets. (Ni, 2024) In 2022, Google's

ExoMiner neural network also screened Kepler

telescope data and discovered 301 previously

unknown exoplanets. (Valizadegan, et al. 2022)

Additionally, AI algorithms can be used to

analyze the spectra of exoplanet atmospheres. AI

algorithms can identify key molecules within them.

For example, by analyzing spectra to search for water,

methane and other molecules that may indicate the

existence of life. Regarding the limitations of

artificial intelligence algorithms in detecting

exoplanets, it is impossible to determine whether

there is any detection bias. But for the increasingly

large artificial intelligence team, it has become the

fastest and most convenient auxiliary tool in life. It

can effectively assist astronomers in exploration, so it

will be a very good method.

6 COMPARISON, LIMITATIONS

AND PROSPECTS

From the entire article, the advantages and limitations

of the three methods can be summarized as follows.

The transit method can perform batch detection, it is

suitable for detecting planets in the habitable zone of

stars, and can directly measure planetary radius, but

the sizes of both the star and planet need to be

considered. The limitation is that it requires

observation of at least one complete transit cycle, and

since the cycle length cannot be predetermined,

researchers need to conduct continuous observations.

It cannot directly measure planetary mass.

On this basis, the radial velocity method can

directly measure planetary mass. However, it is only

highly sensitive to massive planets, its observation

cost is very high, and it is easily affected by stellar

activity. Finally, AI detection is highly efficient, can

prcess complex data, reduce human errors, and

uncover more hidden signals. However, its limitation

is that it relies on extremely precise data and cannot

make independent judgments.

On this basis, this study will propose combining

these three scientific methods to complement each

other. For planets suspected of having signs of life

that have been discovered, the transit photometry

method can be used first, followed by precise

detection and calculation using artificial intelligence,

and finally verified through the radial velocity

method. This process can efficiently confirm

planetary parameters and determine habitability

(David, et al., 2014).

7 CONCLUSIONS

In summary, this is all my views and research on the

three detection methods. Through analyzing the

principles of each method and their currently

observable developments, this research has

conducted self-integration and proposed an

innovative combination of these three different

methods. Recently, more and more research

achievements have been published, and people's

techniques for exploring the universe are becoming

increasingly refined. It is hoped that this can help

promote development, enhance people's emphasis on

cosmic technology research, and stimulate everyone's

innovative consciousness.

REFERENCES

Charles, S. C., 2020 The eventful history and exciting

future of astrobiology. News, 6.

Chauvin, G., 2023. Direct imaging of exoplanets: Legacy

and prospects. Comptes Rendus. Physique, 24(S2),

129-150.

Deng, Z., Tang, Y., 2024. Searching Extra-Planet Based on

Radial Velocity, Transit and Direct Imaging. IAMPA

International Conference on lnnovations in Applied

Mathematics, Physics and Astronomy, 11.

Madhusudhan, N., Constantinou, S., Holmberg, M., Sarkar,

S., Piette, A. A., Moses, J. I., 2025. New Constraints on

DMS and DMDS in the Atmosphere of K2-18 b from

JWST MIRI. The Astrophysical Journal Letters, 983(2),

L40.

Mayor, M., Queloz, D., 1995. A Jupiter-mass companion to

a solar-type star. Nature, 378(6555), 355-359.

Ni, S., (2024). With Innovative AI Technology, Scientists

Discover 5 Ultra-Short-Period Planets. China Science

Daily, 10, 21, 001.

Polanski, A. S., Lubin, J., Beard, C., et al., 2024. The TESS-

Keck Survey. XX. 15 New TESS Planets and a Uniform

RV Analysis of All Survey Targets. The Astrophysical

Journal Supplement Series, 272(2), 32.

Rojas-Ayala, B., 2023. Twenty-five years of exoplanet

discoveries: The exoplanet hosts. Planetary Systems

Now, 71-95.

Valizadegan, H., Martinho, M. J., Wilkens, L. S., et al.,

2022. ExoMiner: a highly accurate and explainable

deep learning classifier that validates 301 new

exoplanets. The Astrophysical Journal, 926(2), 120.

Analysis of Extra-Planet Searching Approaches: Radial Velocity, Transit and AI Algorithm Detection

367

Wang, X., 2022, Exoplanets - Exploring the Starry Sea

Beyond the Earth. International Space Science Institute

- Beijing, 11.

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy

368